Trust layer · public source ledger
Trace the claim before you trust it.
Every tool below has a public evidence row: official pricing, documentation, privacy and rights links where published, the last checked date, verification notes, and unresolved claims.
Official firstVendor sources before summaries or third-party claims.
DatedPricing rows show the last verification date.
Unresolved is visibleMissing rights or policy pages are labeled, not guessed.
On mobile, swipe the evidence table horizontally to view all source columns.
| Tool | Pricing evidence | Policy & rights | Verification notes |
|---|---|---|---|
| AIBase Creative | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via the site form by the vendor on 2026-08-28; editorial entry written independently. AIBase explicitly states it is 'an independent platform … not affiliated with, endorsed by, or an official product of' the underlying model providers (Nano Banana, GPT Image, Seedream, Seedance, Veo, Wan, HappyHorse). Terms of Service §4 (User Content and Generated Content): 'You may use Generated Content for lawful personal or commercial purposes' subject to compliance; user grants AIBase a limited, worldwide, non-exclusive, royalty-free license to host/process content as needed to operate. Outputs are not guaranteed unique, copyrightable, or free of third-party rights. Acceptable Use Policy and Privacy Policy both live and dated. Free trial credits available; ongoing work is credit-based paid. Unresolved: Credits-as-cost model makes monthly spend harder to forecast than flat-rate plans; users should run a real task before subscribing.; Veo 3.1 Pro and Seedance 2.0 are gated to yearly plans / credit packs — Basic and Pro monthly cannot access them.; Because AIBase routes to third-party models, each model's provider's terms and safety layers apply on top of AIBase's own.; No explicit SOC 2 / security claims located on the public site.; Affiliate program live: 30% recurring for the first year, PayPal NET-15 payouts. |
| AIDetector.cx Productivity | 2026-11-03 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Not a FormSubmit submission - an inbound affiliate/partnership offer to partnerships@aidetector.cx, held from 2026-10-06 pending two publication blockers. Editorial entry written independently; no vendor copy reused. Both blockers verified cleared on the live site 2026-10-08: (1) the training-use contradiction is gone - both /detector and /privacy now state submitted analysis text is not used to train or fine-tune detection models on any plan, and the old Enterprise/Professional-scoped opt-out sentence no longer exists; (2) the four named third-party testimonials (Forbes / Stripe / HubSpot / Nature Portfolio) have been removed from /detector. The vendor also removed the SOC-2 Type II footer claim and the 98.7% / 1.2s / 12.4M+ / 142+ marketing figures. Verified the same day from the live pages: /pricing (plans 12/29/79 monthly, 120/290/790 annual, 300/1,000/3,000 credits), /privacy (data controller AIDetector.cx Inc., Berkeley CA; 12-month analysis history retention, deletable from the dashboard; no training use), /terms (you retain ownership of submitted content; output licensed for personal or commercial use; governed by California law), /cookies (Cloudflare / Stripe / Supabase), and a sitemap listing ~40 pages. Operator: AIDetector.cx Inc., a named legal entity, not a sole-proprietor microsite. Name is shared with unrelated live detectors on aidetector.com and aidetector.ai, so the listing carries the exact .cx domain. Affiliate terms confirmed in writing 2026-10-08 - 30% recurring on eligible subscription payments for the lifetime of the customer, monthly and annual eligible, 30-day attribution window, 30-day hold, $50 minimum payout. Unresolved: Payment processor attribution differs across two live legal pages - the Privacy Policy names Paystack for billing while the Cookie Policy names Stripe (plus a Stripe cookie table). Both may be accurate for different layers, but the checkout flow was not exercised.; No independent third-party accuracy evaluation is published; the benchmark figures are the vendor's own results on a 400-document internal corpus.; Security posture is now unanswered rather than contradicted - the former 'SOC-2 Type II Certified' footer claim was removed in the 2026-10-08 fix, and no replacement certification or audit detail is published.; The free tier is five one-time introductory checks (1 guest + 4 signed-in), not a recurring free plan. |
| AILearnHub Education | 2026-10-21 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-20 (13:59 UTC), category 'Writing & Productivity' - reclassified to Education, since the verified product is a lesson generator. Editorial entry written independently. One operator with AIProductAds, TAROTAI.CHAT and BeadPattern; the operator confirmed all four as its own products in a 2026-10-20 15:34 UTC reply. Held 2026-10-20 on no named operator, a product-identity contradiction on its own pages, unreadable contact addresses and legal pages predating the domain. Re-verified live 2026-10-21: the /pricing FAQ still answers 'Free includes all three modes (Bionic Reading / Semantic Bold / Semantic Highlight)' and 'Does Lifetime include cloud TTS?', /privacy still describes browser reading enhancements, and the homepage, nav and /pricing still link Chrome Web Store listing dgihjimekmhphkbnnnomcbemhinmhmeg, titled 'ADHD Reading'. The operator says that extension is a different product; the site continues to present it as this one. No working mailbox: privacy@, legal@ and support@ all return 550 5.1.1 against the Cloudflare MX, and the Gmail printed on /pricing was withheld at the operator's request - so no contact address is published on this listing. PUBLISHED 2026-10-21 on the operator's refusal to name an entity, with 'no registered entity, postal address or named controller' recorded as a stated limitation in cons rather than as a gate, per the standing decision that weak ownership is disclosed rather than used to block. Unresolved: None recorded |
| AIProductAds Video | 2026-10-21 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-20 (14:10 UTC), category 'Video' (accurate). Editorial entry written independently. One of four products from the same operator, confirmed by the operator in writing 2026-10-20 15:34 UTC. Held on no named operator or jurisdiction, an analytics disclosure gap the operator acknowledged in writing, retention with no period, and unnamed model vendors. Re-verified live 2026-10-21: /privacy-policy (dated 2026-10-04) still contains zero occurrences of 'analytics', 'Umami' or 'Clarity' while Umami and Microsoft Clarity (project yi8m85ks1u) load on every page, and the 30:00 'Limited Time Offer / TODAY'S FLASH SALE' countdown still renders on / and /pricing with struck list prices on all four packs. Credit costs are published per resolution. PUBLISHED 2026-10-21 with 'no registered entity, postal address or named controller' and 'no governing-law clause' recorded as stated limitations in cons rather than as a gate. Unresolved: None recorded |
| AccurateScribe.ai Audio & Video | 2026-11-09 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via the site form by the vendor on 2026-11-10 06:30 UTC; editorial entry written independently. Pricing read from the served /pricing HTML on 2026-11-09 and cross-checked against the /ar/pricing locale mirror, which carries the same $9.99 / $119.99 / $19.99. Legal pages /privacy, /terms, /return-policy, /subscription-terms, /copyright, /dmca, /disclaimer and /intellectual-property-rights all returned 200 and were read in full. Operator relationship established from a shared Zoho MX triple, an identical A record, the same registrar, DNS records last changed within 34 hours of each other, 0.90-0.92 whole-body Terms similarity with the two Whisper brands, an identical tool route set, and accuratescribe.com (registered three minutes apart from this domain) redirecting here. Checkout completion not exercised - no account was created. Re-checked 2026-11-09: the vendor's reply of 2026-11-10 14:47 +08:00 was a support auto-responder (canned help on cancelling, refunds, deletion, API, invoices) and answered none of the three submitted questions; it did confirm that website payments settle via FastSpring, that PayPal is used for some website payments, and that iOS subscriptions are billed by Apple. Unresolved: The same operator runs WhisperTranscribe.ai and WhisperTranscript.ai, held rather than listed: the vendor has not said whether this is one product behind three brands or three products, nor which domain is canonical.; No answer yet on the relationship to whispertranscribe.com (live since 2023) and whispertranscript.com, the older .com names the two held brands trade under.; WhisperTranscribe.ai's /subscription-terms publishes a different price ladder (Pro $16.99/1,200 min, $19.99/2,000 min, Business $29.99/6,000 min, no Ultra tier) than its /pricing page (Pro $9.99/24,000 min, Business $19.99/72,000 min, Ultra $73.99/360,000 min); which document governs a purchase made today is unanswered.; The 'Trusted by industry leaders worldwide' band shows Apple, Netflix, Zoom, Microsoft, Fox, Google, YouTube, Meta and Amazon marks; the vendor has not said whether these are customers or illustration.; The 99.8% accuracy figure on the pricing card has no published benchmark, sample size or test set behind it.; The refund policy is absolute and non-refundable; no cooling-off window or UK/EU statutory withdrawal position has been stated.; 'Powered by OpenAI' in the header sits beside an OpenAI mark; whether OpenAI is a paid supplier and what the header is meant to indicate is unanswered. |
| Adobe Firefly Creative | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Adobe's official Firefly FAQ and policy statements say Adobe does not train Firefly foundation models on customer content; current Firefly models use licensed content such as Adobe Stock and public-domain content. Adobe's General Terms say users retain rights and ownership in their Content and Adobe does not claim ownership. The Generative AI User Guidelines state that output may generally be used commercially; if Adobe specifically designates a beta feature as non-commercial, output from that feature is personal-use only. Credits renew monthly and do not roll over. Re-verified 2026-10-01: the four regular monthly plan prices and credit allocations are unchanged. The 30% first-year promotional rates recorded on 2026-08-25 were no longer shown on the pricing page, so only the regular list rates are carried in the summary. Unresolved: The exact free-plan credit allocation was not exposed consistently on the public pricing page and is therefore not stated.; Promotional pricing and unlimited-generation model coverage can change and are governed by Adobe's current promotion page and offer terms.; Adobe does not warrant that output is non-infringing, unique, or copyrightable.; IP indemnification is plan- and contract-specific, particularly for enterprise offerings, and should not be assumed for every Firefly subscription or beta feature. |
| Adobe Podcast Audio & Video | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official plans page checked 2026-08-26. Third-party pages listing Premium at $9.99/month are not treated as official Adobe list prices because that dollar figure was not on the official plans page during verification. Re-verified 2026-10-01: the plans page now attaches a published price to premium access — it is sold as part of the Adobe Firefly Pro plan at US$19.99/month with 4,000 generative credits/month — rather than as a standalone Adobe Podcast subscription. The 2026-08-26 record correctly noted that no standalone price was published, and specifically declined to treat third-party $9.99/month listings as official. Unresolved: Premium access is bundled into Adobe Firefly Pro rather than sold as a standalone Adobe Podcast subscription; the exact entitlement should be confirmed in the Adobe checkout flow.; Daily enhancement caps can change and should be re-read on the live plans page. |
| Airtable AI Productivity | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Airtable's August 10, 2026 billing documentation says AI credits are pooled at workspace or organization level. Credits fund features such as Field Agents, AI automations, deep-research questions, feedback categorization, document analysis, and image generation; consumption varies by input, selected model, and output. The Airtable AI Terms, last updated July 23, 2026, classify Input and Output as Customer Data and state that customers retain all ownership rights in both. Airtable states that neither it nor its third-party AI providers will use Input, Output, or other Customer Data to train generative-AI models. For Business and Enterprise Scale, providers do not log Input or Output for human review or retain it, subject to stated exceptions. For Free, Team, Pro, and Plus workspaces, providers may retain Input and Output for up to 30 days for safety and compliance moderation. Unresolved: The ownership clause does not guarantee that generated output is unique, copyrightable, or free of third-party rights.; Third-party providers may retain metadata for billing, safety, and compliance even where Input and Output retention is prohibited.; The current AI Terms still mention legacy Pro and Plus workspaces in the retention section even though the current public pricing lineup is Free, Team, Business, and Enterprise Scale. |
| AllVideoAI Video | 2026-09-18 Pricing ↗ | Not published Privacy ↗ Not published | Pricing verified from official pricing page 2026-08-23. Submitted via site submission form by vendor (support@allvideoai.com); editorial entry written independently from vendor summary. No rights/output-ownership page published — commercial-use terms should be confirmed in ToS before client delivery. Unresolved: Output ownership/commercial rights not documented on a dedicated page |
| BeadPattern Creative | 2026-10-21 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-20 (14:30 UTC), category 'Creative' (accurate). Editorial entry written independently. One of four products from the same operator, confirmed in writing 2026-10-20 15:34 UTC; its placement also rests on its own published footer cross-links to seedaud.io and productshotai.app. Held on no named operator or processor and a pricing presentation that contradicts itself. Re-verified live 2026-10-21: 'Save 17%' still renders beside the $10.00 to $5.00 (50%) strike, the Subscribe buttons still read 'Alipay / WeChat Pay' above a 'Pay safely and securely with Stripe' footer, and the 'Limited Time Deal / Up to 50% OFF' banner is still present. The 00:00:00:00 countdown recorded on 2026-10-20 did not reproduce on 2026-10-21 - the timer configuration is in the payload but no zeroed timer rendered - so that specific claim is not carried into the listing. PUBLISHED 2026-10-21 with 'no registered company, trading name, postal address or controller identity' and 'no payment processor named' recorded as stated limitations in cons rather than as a gate. PRC governing law and the 90-day log-deletion term are credited in pros. Unresolved: None recorded |
| Bolt.new Development | 2026-10-25 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official documentation says Bolt offers Standard and Max agents; Standard is available on Free, while Max requires a paid plan. Bolt's commercial-use documentation states that all code created with Bolt and StackBlitz is the user's code and may be used for any legal purpose, including commercial purposes. StackBlitz's Terms state that it does not claim ownership of data, content, or other material submitted, published, or transmitted through the service. The Privacy Policy says Bolt may use AI Inputs and Outputs to operate, maintain, and improve the service, including AI performance, reliability, and safety, using aggregated, anonymized, or de-identified data. The Privacy Policy says training/improvement opt-out controls may depend on account type or plan, while third-party AI providers process Inputs to provide services to Bolt under contractual confidentiality and data-protection obligations. Unresolved: Bolt's security page says code and prompts are never used to train foundation models, while the Privacy Policy permits aggregated, anonymized, or de-identified AI content to improve AI performance and describes plan-dependent training controls. The exact boundary between 'foundation-model training' and other model/service improvement is not defined.; The public Privacy Policy does not enumerate which plans receive training opt-out controls.; The commercial-use statement does not provide a non-infringement warranty for generated code. |
| BotLoadout AI Agents | 2026-10-19 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-18 (12:24 UTC), category 'Writing & Productivity' — reclassified to AI Agents, since the product is a library of agent templates rather than a writing tool. Editorial entry written independently. Held 2026-10-18 on three blockers: no privacy notice, no terms and no template licence on a site that collects email addresses. Re-verified 2026-10-19 after the vendor replied: /privacy and /terms both return 200, effective 2026-10-19, naming A13 LLC (Texas LLC, 1101 W 34th St #600, Austin, TX 78705, sponsors@botloadout.com), a commercial-use licence with no attribution and no resale-as-a-library clause, and Texas governing law with Travis County venue. Both are linked in the footer and listed in the sitemap. Count re-verified after the vendor's 'every page now says 97': the sitemap carries 97 /t/ URLs, the library's category counts sum to 97, and the archive holds 388 files for 97 folders. The analytics disclosure matches the shipped markup (Vercel Web Analytics only, no GTM/GA4/Plausible/Clarity). Two vendor claims did not reproduce and are recorded in cons: the per-template 'date added' is not on the template page (only the sitewide 'Updated Oct 14, 2026'), and the newsletter gate is cosmetic — the zip downloads with no signup. Unresolved: None recorded |
| Browse AI Automation | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing page checked 2026-09-01. Browse AI is based in British Columbia, Canada; uses AWS US datacenters and Auth0 for authentication. The Terms of Service provide services on an 'as is' basis without warranties of fitness for a particular purpose. A Trust Center lists SOC 2 Type II, GDPR, and CCPA compliance resources (vendor claims). Credits reset at the end of each billing cycle and do not carry over unless the plan is upgraded mid-cycle. Annual billing provides a 20% discount. Unresolved: SOC 2 Type II, GDPR, and CCPA marks are vendor-published, not independently certified by AIToolsEssentials.; The Terms of Service do not provide an explicit AI-output ownership or commercial-use clause; extracted data ownership depends on the user's right to access the source website.; Premium plan pricing above the $500/month starting point is quote-specific. |
| CCC (Claude Command Center) Development | 2026-10-24 Not published | Docs ↗ Not published Rights ↗ | submitted via form by vendor; editorial entry written independently. Received through FormSubmit on 2026-10-24 20:32 UTC with pricing_url and docs_url both pointing at the GitHub repository and contact_email amir.fish@gmail.com. No privacy_url is recorded because none exists: /privacy, /terms, /privacy.html, /terms.html and /legal/ all 404 on ccc.amirfish.ai, and the parent brand domain amirfish.ai returns 404 for /privacy, /terms, /privacy.html and /terms.html and publishes no sitemap. Product verified by fetching the live site and its subpages (/, /captain/, /receipts/, /watchtower/, /demo/, /vs-heavyweight-ides.html all 200), the repository metadata via the GitHub API (created 2026-04-13, 202 stars, 25 forks, 8 open issues, default branch main, license reported NOASSERTION, pushed the same day as this review) and the raw LICENSE, SECURITY.md, docs/telemetry.md and docs/cloud-relay/SECURITY-AND-PRIVACY.md. The telemetry claim was checked against source (ccc_server/fleet_jobs.py) rather than against the policy page: the payload and the never-sent list are documented in the module, the retired opt-in ping is described in the same comment, and the kill switch is an environment variable. Website analytics were read from the served /analytics.js: PostHog with persistence memory, person_profiles identified_only, and autocapture, pageleave, session replay, surveys, dead clicks, performance, heatmaps and exception capture all disabled, respect_dnt true. Operator identity is a named individual (Amir Fish, copyright holder in the LICENSE) with a personal site at amirfish.ai naming Sunnyvale, California and a LinkedIn profile, plus amir@amirfish.com published there and amir.fish@gmail.com in the submission; the domain amirfish.ai publishes MX (10 inbound-smtp.us-east-1.amazonaws.com), so a domain address is deliverable in principle, but no company entity, postal address or governing-law clause exists. Name collision checked: at least five unrelated GitHub repositories use the same 'claude-command-center' name (nubbymong, Dicoangelo, sorlen008, afstkla, adamclark64) and none is a live incumbent commercial product trading under CCC, so this is a crowded descriptive name rather than a rights dispute. Third-party coverage exists (a YouTube walkthrough, ai-underdog.com, reporank.net, marketgenius.ai, and an entry in andyrewlee/awesome-agent-orchestrators), so the product is not a single unverified page. No hands-on testing was performed: hands_on_status is not_tested and every workflow claim in the listing is attributed to the vendor's own documentation or to the shipped source. Unresolved: None recorded |
| Canva AI Creative | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Canva states that allowance consumption depends on the tool, model, and task complexity; allowances reset monthly. Canva's privacy page says User Content is not used to improve AI-powered features unless consistent with the user's Privacy Settings, which users can review and change. The AI Product Terms say that, as between Canva and the user and to the maximum extent permitted by law, the user owns Output and may use it for any lawful purpose. Output that incorporates or modifies Canva Licensed Content remains subject to the applicable content license. Outputs may not be unique. Re-verified 2026-10-01: Canva Pro is now US$144/year for one person on the annual view, replacing the US$180/year recorded on 2026-08-25. The published AI allowance has also been restated: Free is now described as up to 20 Standard or Premium AI uses (previously recorded as 200 Standard / 20 Premium), and paid plans are expressed as multipliers of the Free allowance rather than fixed Standard/Premium/Ultra counts. Unresolved: Canva prices are localized and may be personalized by billing region; the summary records the official US annual view observed on the verification date.; The official pages do not translate the shared AI allowance into a guaranteed generation count because task complexity changes consumption.; Canva does not guarantee that AI output is unique, copyrightable, accurate, or free of third-party rights.; Commercial use of output containing Canva library content depends on the separate license for that content. |
| Catch AI Agents | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Independently fetched with a browser user-agent on 2026-09-02. Homepage (200) and /security (200) both render a pricing module: “Hire your admin for $99/month”, “$99/month. Billed monthly. Cancel anytime.”, and “Start with a 10-day free trial.” /pricing, /docs, /rights, /privacy, and /terms on www.catchagent.ai returned 404. Footer links Privacy Policy and Terms of Service to app.catchagent.ai/privacy and app.catchagent.ai/terms (HTTP 200 SPA shells without readable policy text in the static HTML). Announcement tweet https://x.com/therealnirs/status/2095134650431586686 was not used for prices. Unresolved: Closed-beta waitlist copy and a public $99/month module appear on the same homepage; it is unclear whether checkout is open to anyone or gated.; Marketing-host /privacy, /terms, /docs, and /rights 404. app.catchagent.ai/privacy and /terms are JavaScript SPA shells, so the full policy text was not independently readable in static HTML.; SOC 2 Type II, CASA Tier-2, and Google Verified claims are vendor statements; the security page says full audit reports are available upon request.; Named testimonials on the marketing site are vendor-supplied and were not independently corroborated. |
| ChatGPT General AI Assistant | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official OpenAI sources confirm consumer and business plan availability. OpenAI's terms state that, as between the user and OpenAI and to the extent permitted by applicable law, users retain Input ownership and own Output. OpenAI's data controls documentation should be consulted for consumer training controls; business products and the API are covered by separate business-data commitments. 2026-09-03 watch: OpenAI released GPT-6 Astra. Official Path to Astra post (https://openai.com/index/path-to-astra/, 2026-09-01) designates Astra as the first model at Critical cybersecurity capability; advanced cyber workflows are limited at launch to trusted testers / Daybreak Blue. Axios reports limited Daybreak Access first, then ChatGPT Plus / Pro / Business / Enterprise and API developers in the coming days. This is not a new ChatGPT SKU. Plus/Pro plan prices were not confirmed as changed on 2026-09-03 and were not re-verified as dollar amounts on this date. Re-verified 2026-10-01: the official pricing page now exposes US-dollar amounts for Go ($8), Plus ($20) and Pro ($100), which the 2026-08-25 check could not retrieve and therefore left unreported. GPT-6 Astra is still presented as an included feature of Pro rather than a separate consumer SKU. Unresolved: Enterprise pricing requires contacting OpenAI sales.; Subscription prices, included benefits, and model limits vary by country, account type, taxes, and promotions.; Astra ChatGPT/API availability timing, any Astra-specific API token prices, and a public system-card URL were not confirmed as a published SKU on the official pricing page. |
| Claude General AI Assistant | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Anthropic's official pricing page states that Team content is not used for model training by default. Its March 16, 2026 consumer-training article says consumer chats and coding sessions may be used when the user allows such use. Consumer and commercial offerings are governed by different terms; the applicable terms should be checked for the account type. Unresolved: The official pricing page does not provide a single all-in Enterprise total because usage is billed separately at API rates.; Sales-assisted Enterprise arrangements may have tailored pricing or contractual terms. |
| CoCounsel Legal Legal | 2026-11-03 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official CoCounsel Legal product page describes Westlaw- and Practical Law-grounded research, analysis, and drafting, plus Word/Outlook/Teams integrations and a statement that customer data is not used to train underlying models. The $415.35/month Westlaw Advantage with CoCounsel Essentials figure was on the plans page on 2026-08-26. On 2026-11-03 that page starts Westlaw Advantage at $288.60/month and Westlaw Edge with AI-Assisted Research at $170.95/month, and it does not show $415.35. Neither figure is a complete firmwide quote. Unresolved: Standalone CoCounsel Legal seat pricing remains sales-configured.; Starting-at Westlaw bundle prices can be personalized by jurisdiction, attorney count, and term.; Vendor SOC 2 / ISO 42001 and no-training statements should be confirmed in the current customer agreement. |
| Copy.ai Creative | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | The official pricing page was published on 2026-08-17 and was checked directly. Copy.ai's official security page, https://www.copy.ai/security, says it does not train its models on individual customer data, share prompts with other customers, or sell customer data. The Terms say Copy.ai does not claim ownership of User Content; they grant compliant users a non-exclusive worldwide sublicensable license to use, copy, modify, sell, create derivatives from, distribute, display, and perform Generated Content for lawful business purposes. The Terms also grant Copy.ai and its licensors a broad license over User Content for providing, operating, securing, and improving their services. Unresolved: Enterprise pricing is not publicly specified.; Workflow-credit cost varies with computational complexity; the official pricing page does not provide one fixed number of workflow runs per plan.; The Terms grant usage rights rather than promising that generated content is unique or copyrightable.; No separate official image-specific rights policy was found; Copy.ai's Terms address Generated Content generally. |
| CrowdWisdomTrading Research | 2026-11-02 Pricing ↗ | Docs ↗ Privacy ↗ Not published | Submitted via form 2026-10-31 12:11 UTC by gilad@crowdwisdomtrading.com (Zoho MX, deliverable). Editorial entry written independently from live pages. Terms name the operator Tsuroni LTD, laws of the State of Israel, Tel Aviv-Yafo exclusive jurisdiction, ILS 400 liability-floor cap; privacy drafted to GDPR/CCPA categories; risk disclaimers throughout ('educational research', 'past performance does not guarantee future results'). No third-party trackers found on any fetched page. Domain registered 2014-02-11 (GoDaddy). '21,576+ pro traders' is a self-published vendor claim, unverifiable. Unresolved: None recorded |
| Cursor Development | 2026-11-03 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | The official Terms state that Anysphere will not use Content to train, or allow a third party to train, AI models unless the user explicitly agrees. The Privacy Policy identifies limited exceptions for security review, explicitly submitted feedback, or explicit training consent. Privacy Mode is available on free and paid individual plans; Cursor says that when enabled it will not train on the user's data and applies contractual and technical controls to model providers. Users retain their rights in Inputs, and Anysphere assigns to users any right, title, and interest it has in Suggestions. Official documentation describes Cursor as a coding agent for understanding codebases, planning and building features, fixing bugs, reviewing changes, and using integrations, MCP, rules, skills, hooks, cloud agents, and CLI workflows. Re-checked 2026-10-13: the standalone Pro, Pro Plus and Teams tiers that the 2026-08-25 check reported are no longer rendered on cursor.com/pricing. The page now shows the unified The X Subscription ladder (Free $0, Plus $8/mo, Premium $30/mo, Super $100/mo, Ultra $200/mo), reported on 2026-10-12 as a bundle of Cursor, Grok and X Premium. xAI's own subscription documentation states that existing Cursor, SuperGrok and X Premium plans continue and that switching is optional, so legacy subscribers are not repriced. Previous digest retained here for reference only: Free with limited agent requests and Composer access; Pro $20/mo; Pro Plus $60/mo; Ultra $200/mo; Teams $40/user/mo; Enterprise custom (checked 2026-08-25). Re-checked 2026-11-03: Free, Plus, Premium, Super, and Ultra prices are unchanged, and Teams is listed at $30+/user/month. Pro and Pro Plus still do not appear. Unresolved: The public Terms assign Anysphere's rights in Suggestions but do not provide a blanket warranty that generated code is non-infringing or expressly guarantee commercial suitability. |
| CustomGPT.ai AI Agents | 2026-10-27 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Pricing, the plan comparison, the add-on list, and the pricing FAQ were read from the served pricing page on 2026-10-27. /privacy and /terms redirect to /privacy-policy/ and /terms-and-conditions/ (both HTTP 200). /docs redirects to https://docs.customgpt.ai/ (HTTP 200). The privacy policy and terms text were not re-read in this pass, so training-use, SOC 2, and GDPR lines on the pricing page stay vendor copy. The affiliate program page at https://customgpt.ai/partners/ was read the same day. The approved tracking URL https://customgpt.ai/?fpr=george-881fd5 returned HTTP 200. No commission tier for this account was stated with the approval. Unresolved: The comparison table says 90+ languages and the FAQ on the same page says 92+. Both figures are recorded; neither was reconciled.; The pricing page says customer data is not used for training and that SOC 2 Type 2 and GDPR are included. The privacy policy text was not read in this pass.; Annual add-on dollar amounts were not in the extracted monthly add-on list. Only the monthly add-on prices are recorded.; Monthly document processing (10,000 Standard / 40,000 Premium) is a different line from documents per agent (5,000 / 20,000). Both stay as printed.; The public affiliate page lists 15%, 18%, and 20% tiers. The approval did not say which tier this account is on. |
| DREAMMMER Creative | 2026-11-09 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by vendor (FormSubmit, 2026-11-09 23:26 UTC, contact support@dreammmer.com); editorial entry written independently from live pages fetched 2026-11-09 Pacific. Live and 200: /, /pricing, /privacy-policy, /terms-of-service, /refund-policy, /contact, /showcase, /tools, /models, /blog, /llms.txt, /sitemap.xml (68 URLs), plus /cookies and /acceptable-use which alias the Privacy Policy and the Terms respectively - there is no separate cookie or acceptable-use document behind either URL, in / or /zh. Domain registered 2026-07-07 (Spaceship, Cloudflare NS kanye/nataly - a pair in no other ledger row), MX Cloudflare Email Routing. Product is real rather than a landing page: the showcase publishes real conversations with full prompt, model and credit cost per card, and those credit numbers match the pricing table's per-model rates. Checkout is live, not announced: /api/config/public reports commerce_checkout_enabled true, paypal_checkout_enabled true, default_payment_provider waffo, google_auth_enabled true, email_auth_enabled false, email_verification_enabled false; the terms name Waffo as merchant of record and PayPal for PayPal payments. No purchase or sign-in was attempted. Operator identity is the gap: no company, trading name or legal entity on any page, only a postal address in the privacy footer (1703, Chengjiaye Estate, Yangmei Road, Bantian Subdistrict, Longgang District, Shenzhen, Guangdong 518129, China); zero occurrences of 'governing', 'jurisdiction', 'arbitration' or 'venue' in the Terms. Retention is absent despite the policy lede promising it ('how long it is kept' appears in the meta description and the opening paragraph; the body has no retention section and 'retention' appears zero times). Analytics: GA4 G-VRDQ11V636 is injected on every page including /privacy-policy and /terms-of-service, the policy body names no analytics vendor (whole-word sweep for 'google', 'analytics', 'posthog', 'plausible', 'clarity' in visible text returns zero) and no consent banner string exists anywhere; the same id appears in /api/config/public as google_analytics_id. Processors are named only by category except Waffo and PayPal. Contact probed, nothing sent: RCPT TO against the domain's Cloudflare Email Routing MX returned 250 for support@, 250 for privacy@ and 250 for a random local part, so the domain is a catch-all and delivery is confirmed while a human reader is not. Name: no incumbent or famous project trades as DREAMMMER (the triple-M spelling returns only auto-generated directory listings on dropyourai, ai-compass, saasworthy and aitoolboard); no collision gate. Unresolved: None recorded |
| DeepSeek General AI Assistant | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | DeepSeek's privacy policy says Inputs and related personal data may be used to improve services and train machine-learning models and that collected personal data is directly processed and stored in the People's Republic of China. Its terms state that users retain any rights in Inputs and that DeepSeek assigns to users any rights it may have in Outputs, subject to applicable law and the terms; the terms expressly mention lawful uses including personal use, research, derivative-product development, and training other models. Re-verified 2026-10-01: The official site still publishes no paid consumer or team subscription; API pricing remains published separately in the API documentation. Unresolved: No official consumer team subscription or enterprise-seat price was found.; Current API prices are model-, token-, cache-, and peak/off-peak-dependent and are not summarized as a single price.; Output ownership does not establish that every output is copyrightable or free of third-party rights. |
| Descript Audio & Video | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Terms last updated February 25, 2025 say Descript claims no ownership of User Content. Inputs and AI-generated Outputs are User Content, and as between the user and Descript, the user owns them to the extent protectable under applicable law. Descript warns that outputs may not be unique or copyrightable and makes the user responsible for third-party rights and legal suitability. Descript may use Inputs and Outputs for model and service development unless the user opts out; its agreements prohibit third-party AI providers from using those Inputs and Outputs to train their own models. Current Help Center guidance is more specific: production models use no Descript user data, internal R&D uses only data from users who opted into sharing, and Enterprise sharing is disabled by default. Descript requires explicit recorded authorization from the depicted speaker before creating a custom voice clone; it refuses deceased, non-consenting, non-authorizing, or artificial-source voices. Consent source: https://help.descript.com/ai-speech/custom-speaker. Data-controls source: https://help.descript.com/account-and-app-settings/data-privacy. Premium Storyblocks media has a commercial-use license under its EULA, but other libraries may require separate clearance; GIPHY assets cannot be used commercially. Unresolved: The terms establish user ownership of protectable Output but do not provide a single universal commercial-license clause covering every generated feature and bundled asset. Commercial use therefore remains subject to third-party media licenses and applicable rights.; The privacy policy's general language allows AI training and service improvement, while newer Help Center guidance says in-house training is opt-in and production models use no user data. The exact default for a particular non-Enterprise account should be confirmed in its current Share Data setting.; AI-generated output may not qualify for copyright protection and may resemble output generated for other users. |
| Dreamlock Productivity | 2026-11-02 Pricing ↗ | Docs ↗ Privacy ↗ Not published | Submitted via form 2026-10-31 13:11 UTC by costelgn@gmail.com; site publishes hello@dreamlock.app (AWS SES MX). Editorial entry written independently from live pages plus the App Store listing (seller named: Andrii Zagoruiko; App Privacy 'Data Not Linked to You'). Policy names the exact subprocessor path (Vercel relay -> Anthropic Claude for text, OpenAI image model from scene description only), states no advertising/training use with API-terms reasoning, and accurately discloses the two analytics scripts (Ahrefs cookieless + GA4) found in markup. Terms: Delaware governing law; Apple IAP billing/cancellation. Registered 2026-07-29 (Name.com); NS Vercel. Unresolved: None recorded |
| ElevenLabs Audio & Video | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official billing documentation and the live pricing page identify the current public plan structure. The Free plan has no commercial license; shared Free output requires ElevenLabs attribution. All paid plans include commercial rights provided the output was generated during the paid subscription, the user holds the necessary underlying rights, and the use complies with applicable law, the Terms, Prohibited Use Policy, and Service-Specific Terms. Paid-subscription output may continue to be used commercially indefinitely. Beta Services are excluded from commercial and production use. ElevenLabs says it uses certain submitted data to improve its models, but any user can disable 'Improve the models for everyone'; the opt-out applies to newly submitted data. Enterprise data is not used for training by default except as necessary to provide the contracted service. The Prohibited Use Policy bans replicating another person's identifiable voice without consent or legal right, harmful impersonation, and deceptive failure to disclose AI generation. Policy source: https://elevenlabs.io/use-policy Unresolved: Some official search-indexed versions of the pricing page showed older Scale and Business figures. The summary uses the live plan values retrieved for the checked date: $299/1.8M and $990/6M.; Commercial rights can vary for individual products under Service-Specific Terms, and Beta Services have no commercial rights.; The data-use opt-out is prospective and does not undo model-training uses completed before the request was processed. |
| Faceloria Creative | 2026-11-02 Pricing ↗ | Docs ↗ Privacy ↗ Not published | Submitted via form 2026-11-02 09:41 UTC (contact field a third-party mail.grokbot.com alias; site publishes support@faceloria.com on Cloudflare Email Routing). Editorial entry written independently from live pages: pricing, privacy (updated 2026-10-26), terms, cookie policy, acceptable use all 200. Processors named in policy (Waffo, Cloudflare R2, Google Sign-In); source photos auto-deleted within 48h per policy. No controller/entity/jurisdiction named; analytics scripts (pvuv.ai Plausible-compatible + plausible.io) load on all pages but policy mentions analytics only generically. Registered 2026-09-14 (Spaceship); NS lars/nola - a pair in no other ledger row. Unresolved: None recorded |
| Fathom Meetings | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | The repository's Free unlimited-recording/transcription claim remains supported. The repository's plan names and prices are outdated: current individual Premium is $20 monthly/$16 annual; current team tiers are Team and Business rather than Team Edition and Team Edition Pro. The official pricing HTML was published on August 20, 2026, and the privacy policy was last updated August 16, 2026. Fathom may use de-identified data derived from meeting content to train, customize, or improve its in-house models, depending on account settings; users can opt out. Fathom says it does not authorize OpenAI, Anthropic, Google, or other third parties to use personal information or meeting content to train their models. Personal-account users retain applicable copyright and proprietary rights in their content. Content in a company Team Account belongs to the company. Fathom uses content as needed to provide the service and through listed subprocessors. The terms require the user to obtain every meeting participant's required consent before each call begins. Official developer documentation at https://developers.fathom.ai/ confirms a public API, OAuth apps, webhooks, and TypeScript/Python SDKs. Unresolved: The Free plan's unlimited recording, transcription, and storage do not establish unlimited use of every advanced AI, team-search, CRM, API, or automation feature.; Exact global API rate limits and plan-specific API entitlements were not stated on the public pricing page. |
| Fireflies.ai Meetings | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | The repository's Free-plan claim of limited transcription credits is outdated: the official page now says unlimited transcription and unlimited AI summaries. The repository's 800-minute Free storage claim is outdated: the current official page says 400 minutes per team. Fireflies states that it does not use meeting content or personal data to train internal or external AI models and contractually prohibits vendors from training on that data. Users retain rights in User Content but grant Fireflies a broad service-operating license. The terms expressly limit use of that license to providing the service and prohibit using User Content to train generative AI models. The account holder must ensure every meeting participant consents to recording. Fireflies documents both platform-dependent opt-in and opt-out flows, advance email notices, in-meeting notices, pause controls, and bot removal. Developer API documentation is separately available at https://docs.fireflies.ai/getting-started/introduction. Unresolved: Do not claim that the visible bot alone supplies legally sufficient consent; Fireflies places responsibility for obtaining legally required consent on the user.; The pricing page lists API access on Free, but exact endpoint entitlements and rate limits should be verified in the API documentation before publication. |
| Gamma Productivity | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | The repository's 400-credit Free allowance remains supported, but those are one-time starter credits rather than a monthly refill. The repository's Plus $10 and Pro $20 figures do not match the current official annual-pricing view, which shows $9 and $18 monthly equivalents. Paid credits roll over only up to twice the monthly allocation; additional credits can be purchased. Individual Free, Plus, Pro, and Ultra workspaces allow Gamma to use content for AI-feature improvement by default, but users can opt out in account settings. Team and Business workspace content is automatically excluded from training, with the control locked. Gamma does not claim ownership of user content. Users grant Gamma a license needed to operate and provide the service. The pricing page also states that content remains the user's after cancellation or downgrade. Public API documentation is available at https://developers.gamma.app/; the pricing page says API access is available on paid plans and consumes AI credits. Approved and wired 2026-09-13: PartnerStack tracking URL on the Visit CTA. Official pricing, docs, privacy, and terms links stay on gamma.app. Editorial score, cons, and methodology are unchanged. Re-verified 2026-10-01: Free, Plus, Pro and Ultra individual prices and credit allocations are unchanged from the 2026-08-25 record. The Team and Business columns did not render, so those figures remain as recorded earlier rather than re-verified. The FAQ still states that paid credits roll over up to 2x the monthly amount and that Free credits are a one-time allocation. Unresolved: The Team and Business plan prices recorded on 2026-08-25 did not render on the 2026-10-01 fetch and were not re-confirmed.; The official page was captured in its annual-billing state. Monthly individual-plan checkout prices were not independently captured and should not be inferred from the annual discount.; Gamma's terms grant a broad training license where training controls are unavailable or left enabled; publications should distinguish individual plans from Team/Business plans. |
| GatsbyAI Video | 2026-10-21 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-22 03:38 UTC (2026-10-21 20:38 PDT); editorial entry written independently from live pages the same working day. Plan prices were read from the rendered yearly cards on /pricing, cross-checked against the site's own llms.txt, and the monthly-billed equivalents ($19.98 / $39.98 / $77.98) plus the four credit-pack prices were read from the shipped plan config inside the served Next.js chunks, because the Monthly and Credit Packs tabs of the pricing page render client-side and carry no figures in the served HTML. Free-tier claims verified from the public credits endpoint (/api/ai-video/credits returns dailyLimit 2 for an anonymous visitor) rather than from marketing copy. /privacy (last updated 2026-10-09) and /terms (2026-10-13) read in full; neither names an operator, jurisdiction or governing law, the privacy policy names no processor and no retention period, and GA4 G-GDTETMSPS7 loads on both policy pages with no consent control. Contact route verified before relying on it: gatsbyai.app MX is Cloudflare Email Routing, support@gatsbyai.app returned 250 in an SMTP RCPT probe while a random local part returned 550. Domain registered 2026-10-01 (Spaceship, Cloudflare NS); no checkout or purchase was exercised. Unresolved: None recorded |
| GearDex Productivity | 2026-10-21 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-22 05:38 UTC (2026-10-21 22:38 PDT); editorial entry written independently from live pages on the site's Pacific working date 2026-10-21. Prices were read from three surfaces that agree: the application/ld+json Offer array on /pricing (0.00, 4.00, 39.00, 19.00, 179.00), the rendered plan cards and the limit-by-limit comparison table on the same page, and /legal/billing which restates US$4/month or US$39/year and US$19/month or US$179/year; the site's llms.txt carries the same figures. Category was reclassified from the submitted 'Writing & Productivity' to Productivity after reading the product pages - it is a gear-inventory and studio-operations app, not a writing tool. Legal pages read in full: /legal/privacy (last updated 2026-10-19, 16 sections), /legal/terms (2026-10-02, 13 sections, California governing law, liability cap of the greater of six months' fees or US$100), /legal/billing (refunds, renewals, California Automatic Renewal Law notice) and /legal/notice; all four return 200 and the Terms' reference to /legal/billing resolves. Privacy section 1 names the controller as Dylan Andersen, GearDex, 842 St Vincent, Irvine, CA 92618, matching /legal/notice and /contact. Trackers were checked on the served pages rather than assumed: Google Analytics G-BNVX42Q1BT loads with Consent Mode defaulting analytics_storage and ad_storage to denied and send_page_view false, and the policy scopes Analytics to accepted cookies, with an Accept All / Reject All banner present in the markup. Contact route: geardex.app publishes no MX (DNS returns only the SOA for the apex), so the address on every page and in the submission - geardex.help@gmail.com - is the working route, not a @geardex.app address. Domain registered 2025-06-28 with Namecheap on Vercel DNS, 16 months old at the check. The privacy policy describes iPhone and Android apps that are not shipped - /download offers Safari Add to Home Screen for iPhone, lists no Android app and carries no App Store or Play link - and that discrepancy is recorded in cons rather than treated as a blocker. Vendor usage claims (300+ photographers, 6,100+ items tracked, $722K+ managed) are unverified by us. No purchase or account was created and no checkout was exercised. Unresolved: None recorded |
| Gemini General AI Assistant | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Google's official Gemini Apps Privacy Hub, updated August 10, 2026, says Gemini data may be used to maintain and improve services and the machine-learning technologies powering them; users can manage whether activity improves Google AI through Gemini Apps Activity, subject to stated exceptions. Workspace editions provide enterprise-grade security and privacy, with availability and limits varying by edition and region. Unresolved: Subscription prices, included benefits, and model limits vary by country, account type, taxes, and promotions.; Enterprise Workspace pricing requires Google Sales or a reseller.; Rights in generated material remain subject to Google terms, third-party rights, and whether applicable law recognizes intellectual-property protection. |
| Genjutsu AI Video | 2026-10-26 Pricing ↗ | Not published Privacy ↗ Rights ↗ | Submitted via form by the vendor on 2026-10-27 00:19 UTC (FormSubmit, contact field harris00525@gmail.com, category "Video", pricing_url https://aigenjutsu.app/pricing, docs_url blank); editorial entry written independently and no part of the submission copy was reused. New lead - no prior genjutsu hit in the repository, the held-submissions ledger or any mail folder. Domain aigenjutsu.app registered 2026-10-08T07:46:49Z at Spaceship, Cloudflare NS paige/derek (a pair that appears in no other row of this ledger). Product is live and functional, not a pre-launch draft: /video renders a working studio with a model selector (Seedance 2.5 default) and a 119-credit estimate for 720p/5s, the homepage motion studio loads presets with labelled photo roles, and /pricing sells three subscription tiers and three one-time packs. Legal set is complete on short paths - /privacy (2026-10-09), /terms (2026-09-28), /refund (2026-10-10), /acceptable-use and /cookie all 200; /privacy-policy, /terms-of-service, /refund-policy, /legal, /about, /contact, /faq, /affiliates, /partners and /referral all 404, and /llms.txt 308-redirects to /sites/aigenjutsu/llms.txt, which is the live copy. Contact route probed against the domain's own Cloudflare Email Routing MX before any mail: support@aigenjutsu.app returned 250 2.1.0 Ok while a random local part returned 550 5.1.1 Address does not exist, so the published mailbox is real and the domain is not a catch-all. Name family - crowded, asked rather than asserted: Genjutsu is Higgsfield's video-to-video model, released 2026-08-31, and seven other domains carry the name (genjutsu.app 2026-10-02, genjutsuai.app and genjutsuai.net 2026-09-30, genjutsu-ai.com and genjutsu-ai.org 2026-10-04, genjutsuai.art 2026-10-05, genjutsu-trends.com 2026-10-11), each on a different Cloudflare NS pair and each serving its own build rather than this platform's data-site template - so no single operator farms the name and no rights demand is warranted. This site discloses non-affiliation in the footer of every page, in llms.txt, on /pricing and on a dedicated /higgsfield-genjutsu page headed "Higgsfield Genjutsu Alternative" that states it does not claim the same engine; the models it actually runs are Kling 2.6/3.0 Motion Control, Wan 2.2 Animate Move/Replace and Seedance 2.0/2.5, all independently confirmed as shipped (Seedance 2.5 launched 2026-07-31; Kling 3.0 Motion Control in February 2026). Published as 3.9 with the identity, retention, training and processor gaps in cons; no affiliate program exists to wire. Terms date 2026-09-28 predates the domain's 2026-10-08 registration - asked, not asserted as stale. Unresolved: None recorded |
| GitHub Copilot Development | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | GitHub's Terms effective April 27, 2026 state that GitHub does not claim ownership of Copilot Input or Output. For individual licenses, GitHub and its affiliates may use Inputs and Outputs to develop, train, and improve AI/ML models unless the user opts out in account settings. The individual-license training provision does not apply when use is governed by a GitHub Customer Agreement or volume licensing agreement; those agreements control data use. GitHub says Inputs and Outputs are not shared with third-party AI model providers for those providers' independent model training. GitHub warns that Output can resemble training data or third-party/open-source code. Users must determine whether a license is required and review, test, and validate Output. GitHub's broader Terms grant GitHub and its affiliates rights to use hosted user content to develop and train AI technologies; the Copilot opt-out does not apply to public-repository content or the broader repository-content license. Re-verified 2026-10-01: The official Copilot plans documentation still matches the recorded Free/Pro/Pro+/Max and Business/Enterprise prices and AI-credit allocations; no drift detected. Unresolved: GitHub does not publish a numeric AI-credit allowance for Copilot Free or Student on the plans page.; The public Terms do not give a blanket commercial-use warranty, non-infringement guarantee, or ownership assignment for generated Output; they only state that GitHub does not claim ownership.; Starting April 22, 2026, new self-serve Copilot Business sign-ups for organizations on GitHub Free and GitHub Team were temporarily paused; the official page did not publish an end date. |
| Grammarly Productivity | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official sources confirm AI drafting, rewriting, tone adjustment, fluency suggestions, grammar assistance, plagiarism checking, and AI-text detection. For Free, Premium, and single-user Pro accounts, Product Improvement and Training is enabled by default, but users can opt out; when disabled, user content is not used for model training, validation, or product improvement. Multi-user accounts bought directly online are also enabled by default, with an administrator opt-out. Enterprise-tier, sales-purchased Pro/Business, and Education accounts have training disabled by default. Grammarly says its external LLM service providers may receive prompts and context to provide generative features but may not train their models on user content. Under the October 29, 2025 Superhuman terms, inputs and outputs are User Content, and users own all right, title, and interest in User Content; the terms also grant Superhuman a broad license for operating, improving, customizing, and developing its services. Unresolved: The billing cadence represented by the $12 Pro price was not rendered unambiguously in the retrieved official pricing text.; Grammarly's exact treatment may differ for negotiated Enterprise agreements, which are governed by a separate Customer Business Agreement rather than the public self-service terms. |
| Grantable Nonprofit | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing page and billing docs checked 2026-08-26. Docs last updated July 30, 2026 describe Free as a real plan and state that canceling moves the account to Free while keeping workspaces. Annual billing is advertised as a savings versus monthly; the public pricing page showed a Save 17% annual toggle, while older docs mentioned 10%. Use the live pricing page at purchase time. Unresolved: Public dollar price for Pro+ was not listed on the main pricing page during this check.; Annual versus monthly savings percentage should be read from the live pricing toggle, not archived copy.; Nonprofit-discount eligibility and Form 990 review are completed by the vendor and may change. |
| Grok General AI Assistant | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | xAI's consumer terms, effective June 26, 2026, state that users retain ownership rights in User Content to the extent permitted by law. Logged-in users can choose whether User Content is used for product improvement and model training; logged-out use may grant broader training rights where permitted. The standalone Grok service and Grok accessed through X are governed by different privacy policies. Re-verified 2026-10-01 against x.ai/pricing and the official model documentation. Consumer plan prices are unchanged, but the API section is superseded: grok-4.7 is now the flagship at $2.00/$6.00 per 1M tokens, and the 2026-08-25 record's Grok 4.6 and Grok 4.3 ($1.25/$2.50) figures no longer appear in the official model overview. The recorded voice figure of '$0.05/min' was also superseded by the published Voice API rates above. Unresolved: Enterprise prices are not published and require a sales quote.; Pricing or entitlements for Grok accessed through an X subscription were not treated as equivalent to standalone SuperGrok plans.; The consumer terms require attribution under xAI's Brand Guidelines when using Output or xAI brand elements; applicability to a specific commercial use requires case-specific review.; Legacy xAI API models (including Grok 4.6 and 4.3) were not re-priced on 2026-10-01; only the currently published flagship rate was verified.; Re-checked 2026-10-13: the SuperGrok tiers above are now also reachable through The X Subscription (a 2026-10-12 bundle of Cursor, Grok and X Premium) at Free $0, Plus $8/mo, Premium $30/mo, Super $100/mo and Ultra $200/mo. xAI documentation states existing SuperGrok plans continue and switching is optional, so the standalone figures recorded on 2026-10-01 remain the standalone list prices. |
| HOL Registry Development | 2026-10-14 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via the site form by the vendor (name "HOL Plugins", url https://hol.org/plugins, category Development, contact michael@hol.org) - FormSubmit mail timestamped Thu, 2026-10-15 00:22 UTC, i.e. 2026-10-14 in the site's Pacific working date. Editorial entry written independently; no vendor copy reused and no revenue figure inferred. Operator identity verified twice: (1) the site's own About and Contact pages state that "Hashgraph Online DAO LLC operates hol.org", give the business address 852 Long Island Rd, Majuro 96960, Marshall Islands, and publish +1 347-705-7394 and support@hol.org; (2) independently, the GitHub organisation "hashgraph-online" (display name "HOL (Hashgraph Online)", 66 public repositories, created 2024-02-18) lists hol.org as its blog, Marshall Islands as its location and support@hol.org as its email, and its hol-guard repository lists https://hol.org/guard as its homepage (Apache-2.0, 896 stars, last pushed 2026-10-15). Submitted contact michael@hol.org is a real mailbox: SMTP RCPT probe against the domain's own Google Workspace MX returned 250 2.1.5 OK, while a random local part returned 550 5.1.1, proving the domain is not a catch-all. Provenance limits, stated plainly: hol.org serves its HTML, its sitemap.xml, its OpenAPI document and its live API endpoints behind a Cloudflare challenge (403 to every scripted client; only the llms.txt surfaces load), and the interactive browser was unavailable on this host, so the catalogue, About, Contact and Privacy pages were read from Wayback snapshots (catalogue 2026-08-31; About 2026-09-19; Contact 2026-09-19; Privacy 2026-09-24) and cross-checked against the live surfaces that do load: /plugins/llms.txt, /guard/llms.txt, robots.txt and the live 404 app shell, which carries the site's full navigation. Ownership is not in question on any reading. Unresolved: No general site-wide Terms of Service was located. The footer link labelled "Terms of Service" resolves to https://hol.org/points/legal/terms (a points-programme legal page) in both the live app shell and the archived pages, and that page's actual scope could not be read because live HTML is Cloudflare-challenged. Treat the recorded rights_url as the only terms link found, not as confirmed coverage of the catalogue.; Catalogue size disagrees between HOL's own surfaces: the archived catalogue page (2026-08-31) read "Showing 243 plugins across 16 ecosystems" while the live /plugins/llms.txt says "205 catalog plugins". The listing therefore quotes no figure.; Retention periods and named subprocessors are asserted in the privacy summary ("the complete policy names categories of providers and explains ... retention") but the full Privacy Policy, stated as last updated 2026-06-29, could not be read from this host. Retention and processor specifics are unverified.; Security scoring is a fallback static catalogue scan, not a live exploit test - disclosed by the site itself in its FAQ and llms.txt. Any trust score should be read as a weak signal, and the listing's pros/cons say so.; The privacy summary warns that information written to a public blockchain can be visible to third parties and may be permanent even after an off-chain account is deleted. This applies to HOL Points and registry features rather than the plugin catalogue, but is not itemised per product. |
| Harvey Legal | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official homepage copy describes Harvey as AI software for legal and professional services, introduces Harvey II, and states enterprise default controls including SAML SSO, audit logs, IP allow-listing, and data lifecycle management. Official customer-count marketing figures are vendor claims, not independently audited here. Re-verified 2026-10-01: Harvey's official homepage still publishes no self-serve price list and routes to a demo request; the only dollar figures on the page are the $550M funding round and $15.5B valuation, which are not prices. Unresolved: No official public list price was published on the homepage during this check.; Data-processing, training, and privilege-safe deployment terms are contract-specific.; Marketing counts of organizations, professionals, and AmLaw 100 usage are vendor claims. |
| Heidi Health Healthcare | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official comparison tables list unlimited transcription on Free and additional Clinician/Teams capabilities. The page states subscriptions can be cancelled and annual plans are not charged for the following year after cancellation. Treat any HIPAA, BAA, or security claim as vendor language to verify in contract review. Re-verified 2026-10-01: the official pricing page now renders US-dollar amounts for the Clinician seat, which the 2026-08-26 check could not retrieve. The page serves locale variants that render different figures, so the US annual view observed on the verification date is the one recorded. Unresolved: The pricing page serves multiple locale variants with different rendered figures; the recorded $110/user/month reflects the US annual view on the verification date.; BAA availability, data-residency options, and EHR integration scope are contract-specific.; A vendor security or HIPAA mark is not independently certified by AIToolsEssentials. |
| Hermes Agent AI Agents | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official documentation describes Hermes as an open-source autonomous agent with tools, skills, memory, profiles, gateways, MCP, scheduling, and multiple providers. Verify current feature names and provider pricing before purchase or deployment. Re-verified 2026-10-01: The official documentation still presents Hermes Agent as MIT-licensed open-source software with provider, hosting and infrastructure costs billed separately; no published product price list. Unresolved: Total operating cost depends on model provider, tool gateway, hosting, terminal backend, and workload.; Feature availability and setup commands can change quickly as the project evolves. |
| HeyGen Video | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Terms last updated July 23, 2026 distinguish plan rights: Creator, Pro, and Business users own their User Input and User Output as between the parties, and HeyGen does not restrict commercial use of User Output; Free-plan Output receives only a revocable license for personal, non-commercial, internal evaluation and may not be sold, monetized, used for advertising, client work, or other commercial services. Terms require AI-origin disclosure where applicable and prohibit unauthorized impersonation or uploading an individual's image without consent. For self-serve paid users, the terms grant HeyGen a royalty-free, transferable, sublicensable, worldwide, irrevocable license to use content to operate and improve services, including training AI/ML models. The privacy policy likewise says inputs may be used for model training and that users may request an opt-out. By contrast, the Enterprise MSA expressly says Customer Data is not used to train AI/ML models and assigns Output ownership to the customer. Official avatar documentation requires proof of consent for a real-person digital twin; webcam consent is generally available, uploaded consent video is Enterprise-only, and skipping the consent flow is limited to Enterprise accounts with an indemnity agreement. Enterprise source: https://www.heygen.com/master-saas-agreement Unresolved: HeyGen's FAQ and some product marketing say user data is not used for model training, but the current self-serve privacy policy and terms expressly permit training and provide an opt-out. These official statements are inconsistent.; The public pricing page exposes many Pro credit tiers but does not provide a compact complete tier table in the retrieved text; only the verified range from 1,000 credits/$49 to 100,000 credits/$4,300 is reported.; Commercial-use language in the self-serve terms expressly names Creator, Pro, and Business; Enterprise rights are governed separately by the MSA and Order Form. |
| HtmlSlides Productivity | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Not published | Submitted via form by vendor; editorial entry written independently. Official site, pricing, terms, and privacy pages verified 2026-08-27. Terms last updated July 11, 2026. Pricing may change — launch offer price noted. Unresolved: None recorded |
| HumanPen Productivity | 2026-10-23 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | submitted via form by vendor; editorial entry written independently. Received through FormSubmit twice on 2026-10-23 21:35 UTC with a byte-identical payload (INBOX msgs 201 and 202 / All Mail 496 and 497) - one submission, not two; the timestamps carry minute resolution only. Submitted contact nitubag316@gmail.com differs from the support@humanpen.net the site publishes; 'HumanPen' had zero hits in the repo, the hold ledger or any mail folder and no prior outbound mail existed to either address. The site is a Next.js SPA, so much of it is client-rendered: /privacy, /terms, /contact, /affiliates and a dozen other guesses return the identical 56,457-byte app shell and are not the legal pages - the real ones are /legal/privacy and /legal/terms (both in the sitemap, both 200 when rendered) and both are dated 'Last updated: 7 September 2026'. Pricing was read from the rendered page AND from the site's own public packages endpoint; the pack cards do not exist in the served HTML at all, so a markup-only price sweep finds just the $19/$9.90 figures in the meta description and would have under-reported the ladder. Privacy policy contents verified by rendering: processors named as Creem and Waffo Pancake (merchant of record, card data never stored), an unnamed authentication provider, an unnamed cloud object-storage provider and unnamed third-party AI model providers; retention 7 days for uploaded documents and results, 30 days for account deletion, 7 years for payment records; no sale of personal data and no training use. Terms verified: user retains rights to input and output, 16+ eligibility, refunds of untouched purchases within 14 days, credits auto-returned on a failed job, acceptable-use clause that explicitly bars academic dishonesty, indemnification, liability capped at three months of payments, and a governing-law clause that names no jurisdiction ('the laws of the jurisdiction in which the HumanPen team operates'). Tracker finding verified from the served HTML and the live network, not from policy text: the root layout injects googletagmanager.com/gtag/js?id=AW-18511954133 and a Microsoft Ads UET loader for bat.bing.net/bat.js, both with next/script strategy 'lazyOnload'; loading /legal/privacy produced requests to googletagmanager.com, googleadservices pagead/viewthroughconversion/18511954133, google.com/rmkt/collect/18511954133 and bat.bing.com/action, and the page contains no cookie/consent UI element and no consent-mode call. Privacy section 10 says the only cookies are those 'necessary to sign you in and keep the Service working' and 'We do not use them to build advertising profiles', and section 5 names no Google or Microsoft processor. Operator identity: none. 'HumanPen' / 'the HumanPen team' is the only party named on any page; no company, individual, country, postal address or company number appears, and the About page explicitly declines to give a biography. Contact route verified before any reliance on it: SMTP RCPT TO against the domain's own Cloudflare MX (route1/route2.mx.cloudflare.net) returned 250 2.1.0 Ok for support@humanpen.net and 550 5.1.1 Address does not exist for a random local part, so the mailbox is live and the domain is not a catch-all. Domain: RDAP registration 2026-07-17T06:52:45Z, registrar Namecheap, NS elaine/rajeev.ns.cloudflare.com, Cloudflare Email Routing MX. No affiliate or partner page is public - /affiliates and /partners both render the sign-in screen - so no tracking URL exists and none was invented. Name checked: no famous research project, open-source repo or incumbent AI product trades as HumanPen; the product's own footprint is a GitHub org (humanpen) with ai-humanizer-comparison, humanpen-skill and humanpen-mcp, an aicatalog.app directory review, a Hugging Face space and a twelve.tools badge. Documented API defect recorded: GET /v1/health and GET /v1/ready are in the vendor's OpenAPI 3.1 schema and both return 404 on the live API. Category reclassified from the submitted 'Writing & Productivity' to Productivity, which is where this catalog keeps writing/document tools. Unresolved: None recorded |
| Ideogram Creative | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing and terms checked 2026-08-28. Paid plans add private generation; Ideogram says it does not claim ownership of generated images and permits commercial use subject to its terms. Re-verified 2026-10-01: Ideogram plan prices and credit allocations (Plus $15/mo billed annually, Pro $42/mo billed annually, Team $20/user/mo billed annually, 1,000/3,500/1,500 priority credits) are unchanged on the official pricing page, which still records paid plans as not claiming ownership of generated images and permitting commercial use subject to the terms. Unresolved: None recorded |
| ImagesV2 Creative | 2026-10-22 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-22 12:18 UTC (05:18 PDT); editorial entry written independently from live pages on the site's Pacific working date 2026-10-22. The submission's summary and category (Creative) were accurate. Prices were read from the rendered /pricing cards and cross-checked against the /zh/pricing mirror and the site's llms.txt; every figure agreed and the per-credit lines reconcile with the credit counts ($49.90/12,000 = $0.0042, $199.90/60,000 = $0.0033, $499.90/240,000 = $0.0021), and yearly allowances equal twelve monthly allowances. Credit costs per image come from the home page pricing block and the FAQ. Legal pages: /privacy-policy and /terms-of-service return 200 with real substance, while /privacy, /terms, /refund-policy, /contact, /legal, /cookie-policy and /acceptable-use all 404; the ToS defers generated-image retention to unspecified 'image storage terms' that are not published. Non-affiliation is disclosed in both the privacy policy and the ToS. Trackers were read from the served markup: Google Analytics 4 G-GC01M2BQCY (gtag config) and Microsoft Clarity (clarity.ms/tag) are injected on /privacy-policy and /terms-of-service themselves, with no consent banner anywhere and no analytics processor named in the notice, against a policy sentence saying data is not transferred to outside parties. Contact: support@imagesv2.ai and alex@imagesv2.ai both returned 250 against the domain's Cloudflare MX with a random local part returning 550 5.1.1, so the published address is a working mailbox and the domain is not a catch-all. Domain registered 2026-04-05 via Dynadot behind a privacy service (Cloudflare NS jake/noor), sixteen days before OpenAI released gpt-image-2 on 2026-04-21 - a pre-launch registration of a model name, not an incumbent under this name. Name family sweeps found four unrelated operators on four different Cloudflare NS pairs (gptimage2.com, gptimage2.ai, gpt-image-2.com, and imagesv2.com now held by an Indonesian gambling site), so the request asks whether a relationship exists rather than demanding rights. No affiliate or partner page exists (/affiliates, /affiliate, /referral, /partners and /api/referrals all 404). Unresolved: None recorded |
| Instrumentl Nonprofit | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing and homepage checked 2026-08-26. Pre-Award copy describes using AI plus past content to write faster. Homepage and FAQ marketing figures such as average grant-revenue increase and hours saved are vendor claims and are not treated as independently verified results. Unresolved: Vendor ROI, hours-saved, and customer-count marketing figures are not independently audited here.; Overage pricing for extra users beyond published seat caps is not listed as a public dollar rate.; Which exact AI drafting features sit on Discover versus Pre-Award can change and should be confirmed in the current trial. |
| Jasper Creative | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pages describe Jasper as a marketing-focused platform with Canvas, Jasper IQ, Brand Voice, Knowledge Base, Audiences, agents, content pipelines, AI Studio, Grid, APIs, and more than 30 supported languages. Jasper's public self-service Terms of Service define inputs and outputs as Customer Property, state that Jasper claims no ownership over Customer Property, and say Customer Property remains the customer's. However, those same terms grant Jasper a royalty-free, sublicensable, transferable, perpetual, irrevocable, worldwide license over Customer Property and state that Jasper uses Customer Property, including inputs and outputs, to provide, maintain, and improve the services, including developing and enhancing Jasper's AI models. Jasper's security page says customer data is retained for the agreement's duration and afterward according to its Data Retention Policy unless a written deletion request is received. The public terms checked were effective May 30, 2025. Unresolved: Jasper's public self-service terms permit model improvement using Customer Property; separately negotiated Business or enterprise contracts may contain different data-use restrictions that are not publicly verifiable.; No universal word, generation, or token limit is stated on the current public pricing page.; The public privacy page does not present a concise AI-training control comparable to the explicit model-improvement clause in the Terms of Service. |
| Khanmigo Education | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official teacher, pricing, and help pages checked 2026-08-26. Teacher-free access is a separate product path from learner/family subscriptions and from district classroom access. Official teacher-page language says teachers cannot grant students Khanmigo access; classroom access is through school or district implementations. Unresolved: Current teacher-eligibility countries and verification rules are completed in the official signup flow and may change.; District Tools pricing is quote-specific.; The number of children covered by a family subscription should be confirmed on the current official parent/pricing pages at purchase time. |
| LM Studio Local AI | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pages describe local model execution, llama.cpp/MLX support, server deployment, SDK options, and optional cloud model billing. Verify current commercial and enterprise terms before organizational rollout. Re-verified 2026-10-01: the official pricing page now publishes dollar amounts for the Bionic+ ($20/month) and Pro ($100/month) cloud tiers; the 2026-08-26 record described 'optional pay-as-you-go cloud models' without figures. Unresolved: The exact boundary between free local features and future commercial/server features can change.; Model licenses and hardware requirements are model-specific. |
| Landlord Studio Property Management | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing page checked 2026-08-26. Vendor savings calculator and review-count marketing are not treated as independently verified results. Per-extra-unit overage dollars were not published as a single public rate on the pricing cards during this check. Unresolved: Exact extra-unit pricing above the Go 3-unit cap is not a single public dollar rate on the pricing cards.; Vendor annual-savings calculator outputs are marketing estimates.; Whether a given bookkeeping feature is branded AI versus rules/OCR should be confirmed in the current product UI. |
| LayerGrab Creative | 2026-11-02 Pricing ↗ | Docs ↗ Privacy ↗ Not published | Submitted via form 2026-10-31 05:55 UTC (contact BlaiseZiemehudfv@hotmail.com; site publishes contact@layergrab.com mailto, MX = Cloudflare Email Routing). Editorial entry written independently from live pages. Chrome Web Store extension verified live (cfkdlnfbbdbihhpbmciigdpbbpeinapa). Privacy names Google sign-in + Chrome Limited Use; uploads 'not sold, not used to train our own models'; the AI model partner is unnamed ('our processing partner'). No entity/jurisdiction/refund page/contact page. Registered 2026-09-27 (Cloudflare registrar); NS eric/reza - not the s07811005141 cluster pair (damon/violet); img2layers.org was unreachable for a text-similarity check, so no cluster claim is made. Unresolved: None recorded |
| Leonardo AI Creative | 2026-08-25 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | The Privacy Policy and Terms were last updated 2026-01-19. Under clause 8.3, paid subscribers own the intellectual-property rights in Content they create, to the extent permitted by law, and Leonardo assigns any corresponding interest it has. Under clause 8.7, rights in free-subscriber Output vest in Leonardo. Paid subscribers can mark Content private; clause 8.5 says Private Content is not used, retained, analyzed, or processed for AI training or new-product development without express written consent. Clause 8.6 gives Leonardo a perpetual, transferable license to Public Content for service operation, promotion, improvement, AI-model training, new offerings, and commercial purposes. Re-check attempted 2026-10-01: leonardo.ai/pricing returned HTTP 403 to every automated fetch route used, so the recorded prices and the 2026-08-25 checked date were left unchanged rather than re-dated on unverified evidence. Unresolved: The reviewed current Terms clearly allocate ownership but do not provide an equally explicit standalone commercial-use license for free-tier users; commercial rights for free-tier output should not be inferred.; Team, enterprise, top-up-token, and API prices vary or require separate pricing paths and are not included in this individual-plan summary.; Unlimited relaxed generation applies only to listed supported models and remains subject to slower queues and fair-use controls.; Leonardo does not guarantee that output is unique, copyrightable, accurate, or free of third-party rights.; The official pricing page returned HTTP 403 to automated fetches on 2026-10-01, so the 2026-08-25 pricing record could not be re-verified; the check must be repeated from a browser session. |
| Lovable Development | 2026-10-25 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official documentation describes Lovable as generating editable frontend, backend, database, authentication, and integration code from natural language, with GitHub synchronization and shared workspaces. Terms effective August 15, 2026 state that users own Customer Data, including applications, websites, and projects, and own AI Output generated for them, subject to third-party rights in underlying models, training data, or outputs. The Terms permit personal or internal-business use, including creating, deploying, operating, and making built applications available to end users. The Terms grant Lovable a broad license to use Customer Data, including for AI/ML development and training, but users can opt out prospectively for free regardless of plan. The official data-control documentation says Free and Pro customer-data training is scheduled to begin September 9, 2026 and users may opt out before or after that date. Business and Enterprise workspace data is excluded from training by default under organization agreements and DPAs. The April 14, 2026 Privacy Policy says raw or identifiable Personal Data is not used for training, while anonymized or aggregated data may be used. Re-verified 2026-10-25: Free remains 5 daily credits capped at 30 a month. Pro is $25/month or $21/month billed annually with 100 monthly credits plus 5 daily build credits. Business is $50/month or $42/month billed annually. Monthly credits still expire two months after issue. Unresolved: As of August 21, the effective Terms already grant training rights with a prospective opt-out, while the product documentation says Free and Pro customer-data training begins September 9, 2026. The sources do not fully explain operational use during the intervening period.; Lovable does not guarantee that AI Output is unique, free of third-party rights, accurate, or suitable for a particular purpose.; Enterprise pricing is not published as a fixed amount. |
| Luma AI Video | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing checked 2026-08-28. Credit costs vary by model and utility; commercial use is listed for paid individual plans. Re-verified 2026-10-01: Official Luma pricing still matches the recorded Plus $30/$300, Pro $90/$900 and Ultra $300/$3,000 plans and credit allocations, with commercial use listed for paid individual plans. Unresolved: None recorded |
| MagicSchool Education | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing and FAQ pages checked 2026-08-26. The privacy page claims FERPA, COPPA, GDPR, 1EdTech, and SOC 2 alignment. Treat those as vendor claims. Free-plan generation limits are described qualitatively on official pages; exact daily generation counts were not consistently enumerated on the pricing page during this check. Unresolved: Exact Free-plan generation caps were not consistently enumerated as a single public number on the pricing page.; Enterprise district pricing is quote-specific.; FERPA, COPPA, and SOC 2 statements are vendor claims, not independent certifications by AIToolsEssentials. |
| Make Automation | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Not published | Official documentation says Make AI Toolkit can analyze sentiment, categorize, identify language, extract, standardize, summarize, translate, and chunk text. Its documentation states that Make does not store submitted AI data or use it to train AI, while separately noting that interaction data remains in standard scenario logs for 30 days. Make's Maia system card says prompts, scenario structures, and outputs are not used to train or fine-tune the underlying LLM; calls sent to an AI provider remain subject to Make's data-processing agreement with that provider. For Make AI Agent, use with Make's provider generally costs 1 credit per operation plus token/tool/embedding usage; custom-provider use generally costs operation credits while the external provider bills model tokens. Re-verified 2026-10-01: the displayed 10,000-credit tier prices (Free $0, Core $12, Pro $21, Teams $38 per month) and the $667 Business tier were confirmed unchanged. The page now advertises saving 15% or more on annual billing. The affiliate page still states 35% commission for 12 months, but a separate answer on the same page says affiliates are paid 'for a period of 24 months' - the two statements conflict and were left as recorded. That page also states Make is 'currently facing delays in processing payouts'. Unresolved: No current official Make AI-specific provision assigning ownership of generated output to the customer was verified; the general Master Services Agreement is therefore not presented as an output-rights guarantee.; The no-training statements were verified for Make AI Toolkit and Maia, but a single universal statement covering every Make AI feature, every connected third-party provider, and every customer configuration was not found.; The AI Toolkit statement that Make does not store data must be read alongside its explicit statement that interaction logs are retained in scenario logs for 30 days.; The official affiliate page states 35% commission for 12 months in one place and a 24-month affiliate commission period in another; the conflict was not resolvable from the page alone.; AI Agent functionality and pricing remain subject to beta-era changes where the applicable documentation labels the feature beta. |
| Makify AI Creative | 2026-11-10 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by vendor; editorial entry written independently. Prices read from the served /pricing HTML on 2026-11-10 and cross-checked against the plan config shipped inside that page: monthly Basic $11.90/100, Standard $29.90/300, Pro slider $59.90-$159.90 for 750-2,300 credits; annual Basic $118.80 ($9.90/mo, 1,800), Standard $238.80 ($19.90/mo, 4,800), Pro $358.80-$958.80 ($29.90-$79.90/mo, 10,800-38,400); one-time packs 80 cr $9.90, 150 $19.90, 250 $29.90, 400 $39.90, 650 $59.90, 900 $79.90, 1,200 $99.90, 2,500 $199 (183 days); subscriber top-ups 300 $29.90, 1,500 $119, 4,500 $349, 8,500 $649, 12,500 $899; Starter Pass 40 cr / 7 days / $4.99. The /zh/pricing mirror carries the same USD figures with a CNY column, and every annual figure equals 12 monthly payments less the badge shown (17/33/50%). Legal set read in full: /privacy-policy and /terms-of-service both 'Last updated Oct 11, 2026', /refund-policy Aug 14, 2026; /privacy, /terms, /about, /contact and /refund all return 404 (the in-app config still points its refund_url at /refund). Sitewide footer names the operator as ARKELUMA LLC; the policy itself names no controller, address or jurisdiction. Whole-word sweep for 'train' across all fetched pages: zero hits. Processors are named only by category. Browser check on a clean profile 2026-11-10: GA4 G-CDGZN0YSBE plus stats.g.doubleclick.net, Microsoft Clarity xtk49ey6wl, c.bing.com and static.cloudflareinsights.com all fired with no consent banner rendered; cookies _ga, _ga_CDGZN0YSBE, MUID, SRM_B, ANONCHK, CLID and localStorage keys device_fingerprint / attribution_session_id were written. MX is Cloudflare Email Routing and accepts every local part, so an SMTP RCPT probe confirms the domain receives mail but not that a human reads a given address; support@makifyai.com is the published contact. Domain registered 2026-04-08 (Cloudflare registrar, NS audrey/sid); the name market is unrelated - makify.com is a DIY/YouTube hobby site and makify.app does not resolve - so no name dispute is recorded. No hands-on test: entering the product requires creating an account, which is outside the scope of an editorial check. Unresolved: None recorded |
| Marigold Pet Video | 2026-10-23 Pricing ↗ | Not published Privacy ↗ Rights ↗ | submitted via form by vendor; editorial entry written independently. Received through FormSubmit on 2026-10-23 15:05 UTC, not through /submit-tool.html but through the site's homepage endpoint, with the submitted URL carrying utm_source=aitoolsessentials&utm_medium=directory&utm_campaign=backlinks - i.e. the vendor is running a directory-backlink campaign and the payload is a link request, not a customer report. Product verified by fetching the live site: /, /pricing, /privacy, /terms and /acceptable-use all 200; sitemap lists 9 URLs including the Spanish mirror /video-halloween-mascotas. Checkout liveness read from the site's own public config endpoint rather than from copy: /api/config returns mode 'prod' and ready true, so this is a live sale, and the client's 'Test checkout is active' string is a dead branch that only renders when mode is not prod - it does not render here and was not treated as a contradiction. Analytics claim verified from the shipped script, not the policy text: consent defaults are denied for analytics_storage, ad_storage, ad_user_data and ad_personalization; GA4 G-V8SX7EYXX3 is appended only on an accepted choice; approved UTM parameters are whitelisted into page_location and order ids stripped; _ga cookies are cleared on reject; a 'Privacy choices' control reopens the banner. Domain RDAP: registered 2026-10-21T13:10:41Z via Cloudflare Registrar, NS adi/theo.ns.cloudflare.com, no MX records. Legal pages are dated 'Effective date: October 20, 2026 - Version 1.0', one day before that registration. Identity: the seller is the brand 'Binglin Studio' only, with no company, individual, country or postal address, and the Terms carry no governing-law clause; the only contact is w1952724273@gmail.com, which is published in the footer and in all three legal pages and is the working route because the domain has no MX. No separate refund page exists (/refunds and /refund-policy 404); refund terms are in /pricing and Terms section 5. Name checked for collisions: no other AI product trades as Marigold Pet (the nearest are an Animal Jam fan blog, a Portland pet-rescue site called Moss & Marigold, and a Claude skill for pet-boarding advice), and 'Binglin Studio' returns no AI-product footprint. No affiliate, partner or referral page exists (/affiliates, /affiliate, /partners, /referral and /api/referrals all 404), so no tracking URL was invented. Unresolved: None recorded |
| Meta AI General AI Assistant | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official sources describe an assistant for questions, planning, actions, and image creation/editing. Meta's Privacy Policy covers AI interactions and related metadata; it states that public information on Meta products and interactions with AI at Meta may be used to develop and improve generative-AI models, with objection rights where applicable. Regional AI terms differ. The general AI terms expressly preserve rights in prompts but do not provide a comparable ownership assignment for outputs. Re-verified 2026-10-01: Official Meta sources still describe Meta AI with Muse Image as free for everyday creation with higher usage via Meta subscription plans, and still publish no subscription names, prices or limits. Re-confirmed that the linked announcement page publishes no consumer subscription price list; the record's no-published-price claim is unchanged. Unresolved: Current Meta AI subscription prices, included quotas, and regional availability were not published in the reviewed official sources.; The reviewed general AI terms do not expressly establish user ownership of outputs or a blanket commercial-use license.; Privacy controls and objection rights vary by jurisdiction. |
| Microsoft Copilot Productivity | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | This verification is scoped to Microsoft 365 Copilot for business, not every consumer or product-specific service branded Copilot. Microsoft 365 Copilot works in Word, Excel, PowerPoint, Outlook, Teams, Loop, and other Microsoft 365 surfaces; it uses Microsoft Graph data the user is permitted to access and includes Copilot Chat, Copilot Search, agents, and app-specific drafting, summarization, analysis, and meeting capabilities. Prompts, responses, and Microsoft Graph data are not used to train foundation LLMs. Interaction content is stored under the customer's Microsoft 365 contractual commitments, encrypted at rest, manageable through Microsoft Purview, and deletable through Copilot activity history controls. Microsoft does not claim ownership of service output but does not determine whether output is copyright-protected or enforceable. Its Copilot Copyright Commitment can defend qualifying commercial customers against certain copyright claims when required guardrails and content filters were used. The official privacy and overview documentation was updated July 9, 2026. Re-verified 2026-10-01: The official Microsoft 365 Copilot pricing page still shows the recorded promotional $18/user/month Business price against the $21 list price, plus the $23.50 and $32 Business Standard/Premium-with-Copilot rates; no drift detected. Unresolved: The general pricing page does not expose one comprehensive price for every Microsoft Copilot product or enterprise licensing scenario.; Pay-as-you-go agent consumption depends on the selected agent and metering arrangement; no single universal usage limit or price is stated on the overview page.; The $18 Microsoft 365 Copilot Business amount is promotional; the checked pricing page does not state its expiration directly. |
| Microsoft Dragon Copilot Healthcare | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official product materials describe physician, nurse, and radiologist experiences and a migration path from Dragon Medical One. Do not copy third-party reseller dollar figures as official Microsoft list prices. Confirm current SKUs and quotes with Microsoft or a certified partner. Re-verified 2026-10-01: The official product page still publishes no self-serve USD list price and routes to a Microsoft representative; Physician Per User and Flex licensing remain documented on Microsoft Learn. Unresolved: No official self-serve USD list price for the Physician Per User license was published on the product page during this check.; Consumption-unit rates and included Flex sessions can change and should be read from current Microsoft licensing guidance.; Availability by country and care setting is not fully enumerated on the marketing page. |
| Midjourney Creative | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | The current Terms are effective 2026-05-27. They say users own Assets they create to the fullest extent possible under applicable law, subject to third-party rights and the Agreement. Companies, and employees of companies, with more than US$1 million in annual revenue must use Pro or Mega to own Assets. Midjourney receives a perpetual, worldwide, non-exclusive, sublicensable, royalty-free, irrevocable license over submitted Content and generated Assets. Content is publicly viewable and remixable by default; Stealth Mode is limited to Pro and Mega. The Privacy Policy covers prompts and uploaded content and says Midjourney may use personal data to provide, maintain, and improve the service. Re-verified 2026-10-01: Official plan pages still match the recorded Basic/Standard/Pro/Mega prices, Fast GPU-hour allocations and the 20% annual discount; no drift detected. Unresolved: The official Privacy Policy discusses training-data collection and service improvement but does not expressly state, in the reviewed text, whether every user's prompts or generated Assets are used for model training.; Midjourney does not guarantee that generated Assets are copyrightable, unique, or free of third-party rights.; Commercial ownership for organizations above US$1 million annual revenue depends on maintaining the required Pro or Mega subscription when creating the Assets.; Relax Mode is unlimited subject to queues and operational limits; it is not a guaranteed throughput rate. |
| Mistral Le Chat General AI Assistant | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official documentation confirms the rename to Vibe and three modes: Work, Code, and Chat. Free-plan conversations may be used to improve models with an opt-out; Pro, Team, and Enterprise conversations are not used for training by default. Connectors and uploaded library documents are not used for training. Consumer terms assign Mistral's rights in outputs to the user, but consumer access is expressly personal; business integrations and products distributed to third parties are directed to Mistral's commercial terms. Unresolved: No blanket commercial-use conclusion was made for consumer subscriptions because the consumer terms limit access to personal use and direct business integrations to separate commercial terms.; Taxes and final localized prices vary. |
| Money Tracker (Expense & Budget) Productivity | 2026-10-22 Pricing ↗ | Docs ↗ Privacy ↗ Not published | Submitted via form by the vendor 2026-10-23 03:33 UTC (2026-10-22 20:33 PDT); editorial entry written independently from live pages on the site's Pacific working date 2026-10-22. The submitted summary's own claim that no privacy page exists is only partly right: /privacy, /en/privacy, /en/privacy-policy and /privacy-policy all 404, but the Privacy Policy served at /Info/privacyPolicy returns 200 and is the exact URL both store listings cite as the app's policy - the marketing site simply never links it (no legal, contact or terms link appears in any of the 45 locales; the footer is "(c) Money Tracker" and no email address appears anywhere on the site). No Terms of Service exists on the domain: /terms, /terms.html, /en/terms, /terms-of-service, /en/terms-of-service, /legal, /agreement, /Info/agreement, /eula, /Info/eula, /Info/terms and /Info/termsAndConditions all 404, while the notice states the terms used have the same meanings as in "our Terms and Conditions, which are accessible at Money Tracker". Notice content verified in full: effective 2023-03-09, contact freemanzk@163.com, third parties named as Google Play Services, Apple Services, AdMob, Firebase Analytics, Crashlytics, Facebook and Adjust, GDPR/CCPA rights sections, no retention periods, no controller postal address or governing law, and no mention of financial information, receipt or bill images, voice recordings, notification content or any AI processor - while Google Play's Data safety declares collection of "Personal info, Financial info and 2 others" and sharing of "Personal info and Device or other IDs", and Apple's App Privacy label (declared by the developer, unverified by Apple) lists Identifiers used to track you and Purchase History, Other User Content, User ID and Device ID linked to the user. The notice also calls the service "an Ad Supported app" "provided ... at no cost" against live in-app purchases, and its CCPA section refers to a "Do Not Sell My Personal Information" section or web page that is not published. Both stores verified as the same product and operator: Google Play com.freeman.moneymanager, 5M+ installs, 4.8 stars from about 324K reviews, updated 2026-10-21, developer account Horoscope365, developer block naming Zheng Kui with a Beijing postal address and support address freemanzk@163.com, website field money.quhou123.com; App Store id6444122615, 4.8 from 414 ratings, seller Zheng Kui, category Finance, Privacy Policy field the same /Info/privacyPolicy URL, in-app purchases $29.99 yearly and $2.99 monthly. Domain quhou123.com registered 2018-02-28 via Alibaba Cloud (Beijing), nameservers vip3/vip4.alidns.com, A record 161.117.61.56 (Alibaba Cloud), and no MX records - so no @quhou123.com mailbox can receive mail and the published 163.com address plus the Gmail used for the submission are the only working routes. The operator's horoscope site www.quhou123.com is live and is the origin of the Horoscope365 publisher name. Name checked: 'Money Tracker' is a crowded descriptive name across both stores rather than a brand dispute, and the App Store description body still calls the product 'Money Manager' while the store title says Money Tracker. Unresolved: None recorded |
| NotebookLM Research | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official about page checked 2026-09-01. The page states: 'We value your privacy and never use your organization's data to train Gemini Notebook.' When used as an individual, data is not used for training unless the user shares feedback. The product is powered by Gemini's multimodal capabilities. A support page (support.google.com/gemininotebook/answer/16179536) provides additional privacy details. Privacy and Terms are governed by Google's policies at policies.google.com. Re-verified 2026-10-01: Google now publishes a dedicated Gemini Notebook plans page, which the 2026-09-01 check on the about page did not surface. The plans page is entitlement-based and still publishes no dollar amounts; pricing sits with the Google One AI Plus/Pro/Ultra subscriptions. The privacy statement is unchanged: organization data is never used to train Gemini Notebook, and individual data is not used for training unless the user shares feedback. Unresolved: The Gemini Notebook plans page does not publish US-dollar prices; the Plus, Pro and Ultra tiers are billed through Google One and are region-limited.; Google's general terms and policies govern use; no product-specific commercial-use or output-ownership clause was found on the Gemini Notebook pages.; The rename from NotebookLM to Gemini Notebook may not be fully reflected across all Google surfaces. |
| Notion AI Productivity | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pages confirm AI features including Notion Agent, Custom Agents, Enterprise Search, AI Meeting Notes, Research Mode, AI writing, translation, file/image analysis, and database creation. By default, Notion and its AI subprocessors do not use Customer Data to train models; subprocessors are contractually prohibited from doing so. Some data-retaining LLM features may be enabled by workspace administrators. Deleted pages/workspaces can be restored for 30 days and are then deleted, including AI-generated data and embeddings. Notion states that it does not claim ownership of inputs or outputs; both are treated as Customer Data. Outputs may not be unique. Supplementary terms were last updated July 27, 2026. Unresolved: Notion does not publicly disclose the exact numerical limits for its rolling six-hour or monthly AI usage allowances.; The exact credit consumption rate for each Custom Agent action varies by operation and is not fully specified on the general pricing page. |
| Ollama Local AI | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official docs state the local API is served at localhost:11434 by default and provide official Python and JavaScript libraries. Ollama's FAQ distinguishes local runs from cloud-hosted models and explains model storage locations. Re-verified 2026-10-01: the official pricing page now publishes dollar amounts for the Pro, Max and Team cloud tiers; the 2026-08-26 record stated only that the page 'lists current cloud tiers'. Official docs still serve the local API at localhost:11434 by default. Unresolved: Each model has its own license and usage terms; Ollama does not make every model legally interchangeable.; Hardware performance, context limits, and output quality vary by model and quantization. |
| Open WebUI Local AI | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official documentation describes Docker, Python, and Kubernetes paths plus connections to Ollama, OpenAI, Anthropic, vLLM, RAG, tool calling, and multi-user capabilities. Read the current license page before rebranding or redistributing. Re-verified 2026-10-01: The official licence documentation still describes self-hosted software with branding and deployment thresholds rather than a published price list; the record's position is unchanged. Unresolved: Backend provider retention and telemetry may differ from the Open WebUI application.; Enterprise features, support, and licensing for larger deployments require direct confirmation. |
| OpenClaw AI Agents | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official documentation identifies OpenClaw as a multi-channel gateway for AI agents. Treat token, hosting, gateway exposure, and extension security as separate procurement questions. Re-verified 2026-10-01: The official documentation still describes OpenClaw as an MIT-licensed multi-channel agent gateway with no published price list; the record's no-price claim is unchanged. Unresolved: The total cost of a deployment varies with provider, model, channel count, and hosting.; Managed-service availability and enterprise features may change; verify the current official offering. |
| OpenEvidence Healthcare | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official homepage copy lists content agreements or collaborations with NEJM, JAMA, Nature, NCCN, and Cochrane, and displays a HIPAA Compliant mark. Eligibility, verification, advertising, and geographic availability should be confirmed in the official signup flow. This is a knowledge/reference product, not a clinical-decision device certification. Re-verified 2026-10-01: The official site still states the product is free and unlimited for verified U.S. healthcare professionals, with no self-serve paid price list; no drift detected. Unresolved: Current professional-verification rules and geographic eligibility are completed in the official signup flow and may change.; The homepage HIPAA mark is a vendor claim, not an independent audit result published here.; Whether answers may be used in a specific clinical or billing context is an organizational and professional-liability question, not a product-page fact. |
| Otter.ai Meetings | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | The repository's Basic allowance of 300 minutes per month and Pro limit of 1,200 minutes are still supported. The repository's Business price of $30/user/month is the monthly rate, but the official annual rate is currently $19.99/user/month. Otter's official Privacy & Security page says it automatically trains proprietary models on de-identified recordings and transcripts; the data is encrypted and not manually reviewed as part of that automatic method. A separate Feedback and Training setting allows users to permit or revoke Otter and third-party-provider access to conversations for training and product improvement, including human review. Selecting 'Don't allow sharing' removes access to past conversations and excludes future conversations under that setting. Otter states that its external AI service providers do not train on or retain customer data sent through their APIs. Customers retain ownership of User Content and grant Otter a service-operating license. Workspace or organization administrators may own and control data placed under an organizational account, including migrated pre-existing data. Users are solely responsible under the terms for legally required recording notices and consent. Otter also provides pre-meeting notices, in-meeting announcements, and an Enterprise Microsoft Teams affirmative-permission flow. Unresolved: Otter's official materials describe both automatic de-identified model training and settings-based training permissions. Do not reduce this to a blanket 'no training' or blanket 'opt-out' claim without clarifying which processing path and account tier is meant.; The pricing page displayed temporary Business promotional pricing alongside standard monthly and annual rates; promotional figures should not be presented as durable list pricing. |
| PawHeadshots Creative | 2026-11-02 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form 2026-11-03 02:39 UTC (contact field a third-party mail.grokbot.com alias, vban@) - editorial entry written independently from live pages fetched 2026-11-02 Pacific. Live: /, /studio, /me-and-my-human, /pet-resume, /resume, /breeds, /privacy, /terms all 200; /pricing, /about, /contact, /faq, /cookies, /refund 404. Operator named as Vban LLC in both policies; processors named by company (Cloudflare, fal.ai running OpenAI and Google models, Stripe and Link as merchant of record); retention stated as periods (uploads 3 days, images 7 days, IP counters 1-2 days, order records for tax); training use explicitly denied; commercial rights granted with the pack and revoked on full refund; 18+ rule. Analytics: Cloudflare Web Analytics is the only tracker on the fetched pages and the policy names it as cookieless - disclosure and instrumentation match. Checkout is wired, not announced: /api/checkout is a live POST route in the shipped client bundle and /api/orders answers GET 200 with an empty list; no purchase was attempted. No governing-law clause, jurisdiction or postal address anywhere (whole-page sweep of /terms for 'governing law', 'jurisdiction', 'courts', 'arbitration', 'address' returns zero). Domain registered 2026-10-30 14:25 UTC (Cloudflare registrar, NS johnathan/athena - a pair in no other ledger row); the name appears on a March 2026 .com drop list, so this is a re-registered generic pet-name domain, not a squat. Contact support@pawheadshots.com verified deliverable before any send: SMTP RCPT TO against Cloudflare Email Routing MX returned 250 2.1.0 Ok while a random local part returned 550 5.1.1, so the domain is not a catch-all (nothing sent on the probe). Unresolved: None recorded |
| Perplexity Research | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Perplexity is documented as an AI-powered answer/search product with citations, premium models, file uploads, Research, Create files and apps, Comet, and Computer features depending on plan. AI Data Retention is enabled by default for Free, Pro, and Max users and can be turned off for future collection; previously collected training data is not removed by a later opt-out. Enterprise data is not used for AI training, and Perplexity says its third-party model-provider agreements prohibit provider training on Perplexity data. Consumer terms warn that output may be subject to third-party model licenses and use restrictions. 2026-09-02 official blog “Optimizing On-Device Inference for Apple Silicon” (https://www.perplexity.ai/hub/blog/optimizing-on-device-inference-for-apple-silicon): Lily is a lightweight local inference engine built specifically for Apple silicon and Qwen3.6-35B-A3B so on-device compute does not bottleneck Hybrid Compute / Perplexity Computer tasks. Official facts from that post: Rust runtime; OpenAI-compatible chat-completions API; custom Metal kernels; neither PyTorch nor MLX in the execution path; standalone demo on GitHub. The official pplx-garden README names the demo at https://github.com/perplexityai/pplx-garden (lily/), Apache-2.0, greedy generation for the Qwen3.6-35B-A3B MLX affine 4-bit checkpoint format, Apple GPU family 10 or later (M5 and newer). Lily is not a new paid Perplexity plan SKU. Consumer Pro/Max dollar prices were not re-verified on 2026-09-02. Re-verified 2026-10-01: the official plan-comparison article still routes consumer Pro and Max to the app rather than publishing dollar amounts, and lists Education Pro at $10/mo with SheerID verification. Enterprise Pro is stated as starting at $40/month or $400/year per seat, matching the 2026-08-25 record. Unresolved: Current public prices for individual Pro and Max were not verifiable from the reviewed official pages.; A blanket commercial-use or output-ownership claim was not made because consumer outputs may be governed by third-party model licenses.; Enterprise Max pricing was not publicly stated in the reviewed sources.; The official Lily blog says the chat-completions API streams tokens; the official lily README in pplx-garden as of 2026-09-02 documents stream: false and rejects streaming. Treat serving-surface details as demo-specific until re-checked.; Lily hardware support is stated as Apple GPU family 10 or later (M5 and newer) and macOS 26 or later in the official demo README; support on every later Apple chip or OS was not independently verified. |
| Pika Video | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Live re-check 2026-09-18: the official pricing route DOES publish a self-service plan table (Starter/Creator/Fancy with a monthly/yearly toggle). The previous record, which stated the page exposed no plan table and carried Basic/Standard/Pro plan names, was wrong and has been replaced with figures read off pika.art/pricing today. Recheck before purchase; Pika reprices and renames tiers. Unresolved: None recorded |
| Poe General AI Assistant | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Poe provides one interface for third-party text, image, audio, and video bots, custom bots/apps, group chats, and shared subscriptions. Fully shaded privacy-shield bots share data with model providers but interactions are not used by those providers for training; half-shaded bots/apps may share data with developers or external resources and may use interactions for training. Poe's terms define prompts and bot-generated results as 'Your Content' and state that users retain ownership, while granting Poe a perpetual license for stated service purposes. Use also remains subject to policies of the underlying model providers. Unresolved: Monthly-billing prices were not included in the extracted official pricing page, so only verified annual prices and their displayed monthly equivalents are reported.; No blanket commercial-use assurance was found because Poe grants a personal right to use the service and underlying model-provider policies may impose additional restrictions.; Privacy and training treatment varies by each bot or app's privacy-shield status. |
| RapVideo AI Creative | 2026-10-27 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-27 14:13 UTC; editorial entry written independently from live pages on the site's Pacific working date 2026-10-27. The submission's name, category (Creative) and summary were accurate. Prices were read from the served /pricing HTML and cross-checked against the /zh, /es, /fr and /de mirrors, which carry identical figures; per-render credit costs come from the same page's model table. The yearly toggle is JavaScript-only, so only the Basic yearly figure ($299/120 videos, quoted in the studio's own price line on the homepage) could be verified - Pro and Agency yearly prices are recorded as unavailable rather than guessed. Three internal pricing inconsistencies were verified on the live pages and are published in cons: 'About $1 a song' (homepage value card plus the /pricing <title> and og:title) against 'about $2.99 a song on the Basic plan' (hero and FAQ) and the $3.32/video Label pack; '100 credits / month · 8 videos' on the Basic card against '10 songs a month for $29.90'; and plan/pack sizing at ~12 credits a video against the stated 10-credit standard render. Legal set: /privacy-policy, /terms-of-service and /refund-policy all 200 (all dated 'Last updated: 2026-10-04'), while /privacy, /terms, /refund, /about, /contact and /faq 404 - the FAQ is the homepage anchor /#faq, not a missing route. Processors named by company with purpose (Kie.ai/ByteDance Seedance, OpenAI lyrics, Cloudflare, Railway, Stripe, Google); training use explicitly excluded; likeness, minors and prohibited-use clauses present; refund windows stated; non-affiliation with artists and labels stated in the ToS and on the Hotel Lobby guide. Trackers read from the served markup on the policy pages themselves: GA4 G-GLWNKQXXM1 via googletagmanager, Microsoft Clarity yr9wgy5s80, PostHog, plus the Cloudflare beacon - none named in the notice, no cookie section, no consent gate (the 'consent' strings in the bundle are photo-permission i18n, and the 'countdown' strings are OTP resend labels, not a pricing timer). No company, controller, postal address or governing-law clause exists in the ToS. Contact: support@rapvideoai.com published via Cloudflare email obfuscation on every legal page and probed live against the domain's Cloudflare MX with nothing sent - 250 2.1.0 Ok, random local part 550 5.1.1, so one real mailbox on a non-catch-all domain; the submission's alex@imagesv2.ai also answers 250 and is the same address that submitted ImagesV2. Operator: rapvideoai.com and imagesv2.ai share the Cloudflare nameserver pair noor/jake, the same Cloudflare MX triple and registrar Dynadot, and the submitted contact address is on the imagesv2.ai domain we already list - same operator, separate product, so this is a second listing rather than a duplicate. Name: rapvideo.ai is a separate operator (Roman/Bristol nameservers, Zoho MX, Spaceship registrar, registered 2026-10-02) selling a same-named Hotel Lobby rap-video product; homepage body-text similarity to this site measured 0.156 with difflib, i.e. separately drafted, and rapvideoai.com was registered 2026-09-29, three days earlier - a crowded fresh name, not a squat, and it is disclosed in cons. Unresolved: None recorded |
| ReelTada Video | 2026-11-03 Pricing ↗ | Not published Privacy ↗ Rights ↗ | Submitted via form 2026-11-02 09:41 UTC by support@reeltada.com (MX = Cloudflare Email Routing); summary self-described as 'suggested by ReelTada's AI assistant'. Editorial entry written independently from live pages. Legal set complete (privacy, terms, refund; AUP-grade content rules in terms) but refund policy carries a leftover template brand ('Bby Wow credit packs') - cited to the vendor. No operator identity/jurisdiction; privacy names no processors/retention/training stance; three analytics hosts (Plausible, CF Insights, self-hosted analysis.gaolong.site, Dnspod NS, registered 2026-10-12) load sitewide unnamed. Domain registered 2026-10-20 (Spaceship); NS blakely/nolan - a pair in no other ledger row. Already listed on third-party directories (aicoming.net, getai.business, withaitools.com, indietools.app). Unresolved: None recorded |
| Replit AI Development | 2026-10-25 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official documentation describes Replit Agent as turning plain-language requests into apps, designs, slides, and other artifacts. The Commercial Agreement states that commercial customers retain all rights in Input Content and own Output Content. Under the Commercial Agreement, Replit will not use Customer Content to develop or improve products, train machine-learning models, or create derivative works except as expressly permitted; customer-requested tuning may be performed solely for that customer's use during the agreement term. The consumer Terms state that users retain their rights in content they submit and grant Replit the operational rights needed to provide the service. Replit's Privacy Policy was last updated August 3, 2026 and covers collection of usage, interaction, project, command, payment, and communication data, along with applicable access, deletion, correction, appeal, and sale/sharing opt-out rights. Unresolved: The explicit Output Content ownership and no-training language is in the Commercial Agreement. The general consumer Terms do not publish an equally explicit assignment of rights in AI-generated Output.; The pricing page labels free and included Agent allowances qualitatively rather than publishing exact numeric Agent-credit limits.; Third-party AI services are excluded from the Commercial Agreement's definition of Replit AI Features and can be governed by separate provider terms. |
| Retouchia Creative | 2026-10-22 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-23 00:11 UTC (2026-10-22 17:11 PDT); editorial entry written independently from live pages on the site's Pacific working date 2026-10-22. Submitted summary and category (Creative) were broadly accurate, though the site is a multi-model photo editor rather than only a background-removal/upscale tool. Prices read from the rendered /pricing cards and the SKU data in the pricing chunk (basic_one_time, pro_one_time, currency usd) and cross-checked against the /zh/pricing and /zh-tw/pricing mirrors and llms.txt - all three agree on $9.99/350 and $29.99/1,200, on the $12 and $39.90 anchors, and on 1-year credit validity, and the stated 16% and 24% discounts match those anchors. The ~70 and ~240 image figures are hardcoded display strings in the pricing bundle, not computed, and they imply 5 credits per output while the Terms state 4. Purchasability confirmed without transacting: Waffo checkout reports enabled in the public config endpoint (/api/config/public returns waffo_enabled true, app_name Retouchia, google_analytics_id G-G107GQMKL9, email_verification_enabled false, password_reset_enabled false), and the packs carry Product/Offer JSON-LD with price, currency, SKU and seller. Operator identity, jurisdiction and governing law are absent by the vendor's own statement rather than by omission - this website does not publish a legal entity name, registered office or postal address. Legal set verified live: /privacy-policy, /terms-of-service, /acceptable-use, /refund-policy and /cookie-policy all 200 and versioned by date. Model provenance is the unresolved item: the sitewide navigation and /gpt-image-2-5 advertise GPT Image 2.5 as "OpenAI's" model, while Terms section 2.1, the AUP and the Privacy Policy name only SparkPix and Nano Banana 2, and the strings GPT and OpenAI appear in no legal page; no non-affiliation statement exists anywhere on the site. Contact probe: contact@retouchia.com returned 250 against route1.mx.cloudflare.net with a random local part at 550 5.1.1 (real mailbox, not catch-all). Domain registered 2026-08-30 via Cloudflare, nameservers dayana/seth.ns.cloudflare.com. Name checked for collisions: no other product trades as Retouchia; directory listings exist on FutureTools, tinylaunch and a wiretensors review dated 2026-09-27 which notes public adoption appears minimal - relevant to the site's unverified 200K+ users / 10M+ images claims. Unresolved: None recorded |
| Riverside Audio & Video | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official riverside.com/pricing and riverside.fm/pricing checked 2026-08-26. Some Webinar cards also display 'Billed $408 annually,' which conflicts with the official $79/month annual equivalent. The $79/$99 Webinar pair from the compare table and FAQ is recorded; the $408 Webinar annual total is unresolved. SOC 2 and ISO 27001 marks on the Business tier are vendor claims. Re-verified 2026-10-01: The official pricing and FAQ pages still match the recorded Free/Pro/Grow/Webinar rates and separate-track download caps; the Webinar $408 annual-total discrepancy remains unresolved. Unresolved: Webinar annual-total copy on some cards shows $408 while official FAQ/compare list $79/month billed annually.; Editor-seat add-on dollar amounts are not published as a public list price.; AI Credit pack prices after the published starter allotment are not listed as a public dollar rate. |
| Robylon AI Agents | 2026-10-16 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via the site form by the vendor (name "Robylon", url https://www.robylon.ai, category "Automation", contact listings@robylon.ai) - FormSubmit mail timestamped Sat, 2026-10-17 04:03 UTC, i.e. 2026-10-16 in the site's Pacific working date. Editorial entry written independently; the vendor's summary was not reused and no revenue figure was inferred. Category reclassified from the submission's "Automation" to AI Agents, which is where the verified use case sits. Operator identity corroborated four ways: (1) /llms.txt and /about state "Operated by Zooid Tech Inc., San Francisco, California"; (2) /about gives the legal entity as "Zooid Tech Inc., incorporated in California", the headquarters as 128 King Street, San Francisco, CA 94107, USA, and the phone as +1 (415) 318-0975; (3) the same page names the two co-founders, Dinesh Goel (CEO) and Mayank Shekhar (CTO), with LinkedIn profiles; (4) independently, the domain's RDAP registrant, administrative and technical contact is "Dinesh Goel" and robylon.ai was registered 2024-01-24 through Namecheap - the registrant name matches the named CEO. The submitted contact listings@robylon.ai is a real mailbox: the domain resolves to Google Workspace MX (aspmx.l.google.com and four backups) and its SPF record includes _spf.google.com and _spf.salesforce.com. Nameservers are Cloudflare matias/addyson - a pair that appears in no held cluster in admin/held-submissions.md - and nothing about this submission resembles the trend-microsite clusters there. Product verified from the live site, its llms.txt (30 KB), its Markdown twins and its sitemap index (508 URLs): four channels, nine named helpdesk/CRM integrations, a knowledge base with gap clustering and a bulk answer-grading suite, a model selector over OpenAI, Anthropic and Google models with fallback, and optional human seats. Terms verified live at /terms-of-service: governing law is the State of Delaware with venue in New Castle County, refund terms are stated in their own section, output ownership and input rights have their own sections, and deletion occurs 30-60 days after termination. Privacy Policy verified live at /privacy-policy: the controller is named in the notice itself (Zooid Tech, Inc.), Paddle is named for payments, and five service providers are named by company in the Terms' Service Providers section (AWS, OpenAI, Firebase, Qdrant, Google Analytics). Provenance limits, stated plainly: the Privacy Policy is dated 7 April 2024 and has not been refreshed for the current product; the trust centre cited on /about (trust.robylon.ai) returns HTTP 403 to a standard browser user-agent, so the SOC 2 Type II and HIPAA claims could not be read; the paths /privacy, /terms, /contact, /security, /trust, /status, /dpa, /subprocessors and /refund-policy all return a Cloudflare 403 ("CNAME Cross-User Banned") while /privacy-policy and /terms-of-service return 200 and are the paths the site itself links; and the general web-search backend returned unrelated results for every query about the company, so no third-party corroboration beyond the site's own G2 profile (5.0 from 42 reviews, read October 2026) was obtained from that source. Unresolved: SOC 2 Type II and HIPAA compliance are stated on /about and in /llms.txt, but the trust centre cited as the evidence surface (trust.robylon.ai) returns HTTP 403 to a standard browser request. The claims are vendor-stated and were not verifiable during this check; the listing's cons say so and no compliance claim is repeated as fact.; Training use on customer data is not addressed directly. The Privacy Policy permits the use of "non-identifiable information, including anonymized, aggregated, or de-identified data, for lawful purposes" but does not state whether submitted prompts or support conversations are used to train models. Robylon's robots.txt notes that blocking "training" crawlers "would also block Googlebot, which is multi-purpose" and allows named AI crawlers - a crawler policy, not a statement about customer data.; International transfers rest on consent ("By using the Service, you consent to the transfer of such information") rather than a named mechanism. No Standard Contractual Clauses or Data Privacy Framework basis is referenced.; Retention is two-tier and partly open-ended: customer and user information is deleted 30-60 days after termination or on request, but usage data is kept only "as long as needed" with no period. Custom retention policies are an Enterprise-only feature.; The Privacy Policy's last-updated date (7 April 2024) predates the current product surface, including voice, WhatsApp and the browser-automation section of the Terms, so parts of it may describe an earlier service.; Enterprise is quote-only ("Let's talk") and per-conversation cost depends on which model answers, so no single headline price exists for the top tier. |
| Rows Productivity | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing page checked 2026-09-01. Rows is operated by Rows GmbH, Berlin, Germany. Terms of use dated June 5, 2023 note that services are offered to entrepreneurs or businesses. Rows joined Superhuman; the terms page footer references superhuman.com/legal/terms. The Terms grant Rows a royalty-free, worldwide, transferable, perpetual license to use user feedback. The Privacy Policy (February 3, 2025) describes processing of Customer Account Data, Customer Spreadsheet Data, Partner API Integration Data, and Stored Integration Data under GDPR. Unresolved: The Terms of use are dated June 2023 and the Privacy Policy February 2025; current AI-feature-specific data-use practices may differ.; The Terms do not provide an explicit AI-output ownership or commercial-use clause.; Enterprise pricing above the published tiers is quote-specific. |
| Runway Video | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing and Help Center pages agree on the 625/2,250/9,500 monthly credit allocations. Terms last updated May 11, 2026 state that Runway does not claim ownership of Inputs or Outputs and does not restrict commercial use of Outputs, subject to compliance with the agreement. The same terms grant Runway a perpetual license to use Inputs and Outputs for model training, labeling, classification, moderation, and related improvement. The privacy policy says submitted voice and face characteristics may be biometric data and are used only to provide the requested biometric-enabled feature. Runway's official voice-verification documentation requires explicit, in-app consent from the same speaker before custom-voice training and prohibits non-consensual replicas. Terms also prohibit impersonation and submitting another person's photograph without permission. Official voice-consent source: https://help.runwayml.com/hc/en-us/articles/29828787723411-Voice-verification Unresolved: The self-serve terms authorize training on Inputs and Outputs, while Runway separately advertises no training on Enterprise customer data. The precise Enterprise exceptions depend on the separate Enterprise agreement and cannot be generalized to self-serve plans.; The pricing page does not provide a fixed Enterprise price or credit allocation. |
| Shortwave Productivity | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing page checked 2026-09-01. Terms of Service last updated January 26, 2023; Privacy Policy last updated July 10, 2025. Terms state users own their User Content and grant Shortwave operational rights. Shortwave integrates with a sister product, Tasklet, for 24/7 workflow automation. The pricing page does not list a permanent free plan; only a 14-day free trial is advertised. Unresolved: The Terms of Service are dated January 2023 and may not reflect current AI-feature-specific data-use practices.; The pricing page does not publish exact API rate limits or team-plan entitlements beyond the listed AI usage tiers.; No explicit AI-output ownership or commercial-use license clause was found in the retrieved Terms; users retain rights in User Content but commercial use of AI-generated output is not separately addressed. |
| Slack AI Productivity | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Not published | Official pages list capabilities including Slackbot, conversation and thread summaries, daily recaps, AI search, file summaries, workflow generation, huddle notes, translation, and enterprise search, with availability varying by plan. Slack states that Customer Data never leaves Slack-controlled infrastructure and is never used to train LLMs. Third-party LLMs are hosted within Slack's cloud environment; providers do not have access to Customer Data. AI only uses content the requesting member is authorized to access. Conversation summaries and search answers are ephemeral; recaps are stored for 90 days. Workflow-generated summaries stored in messages or canvases follow the organization's retention settings. Admins can disable or restrict native AI features. Unresolved: Slack's AI Principles page says AI features are included in paid plans, while the current pricing page labels the Free plan as having Basic AI. The exact production availability and limits of Free-plan AI could not be reconciled from the official pages.; Slack does not publish a single numeric usage allowance for all AI features; some capabilities state that plan limits apply.; No explicit, AI-output-specific ownership statement was located in the checked official Slack sources. |
| Sonaopus Productivity | 2026-09-13 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via /submit-tool.html by the vendor on 2026-09-09 and held pending verification. The vendor published /privacy and /terms on 2026-09-12 and answered every outstanding question, so the listing was published 2026-09-12 with remaining caveats recorded in cons and unresolved_claims. Editorial entry written independently; no submission copy was reused. Verified on 2026-09-12: both legal pages live and substantive; all five Gumroad products at $99 with no subscription; shipped JavaScript writes only sonaopus_hr/_ea/_pm/_re/_role to localStorage, makes zero fetch/XHR/sendBeacon calls, and has no third-party script tags; handoff targets claude.ai/new, chatgpt.com/ and gemini.google.com/ carry no query parameters. Operator: Praise Boyinde, sole proprietor, no registered entity, governing law Nigeria. The domain publishes no MX records, so the Gmail address printed in both legal pages is the only reachable contact. Re-checked 2026-09-13 against the live homepage and a vendor email the same day. Closed: the homepage FAQ no longer says “Nothing stored”; on-page copy now matches the Privacy Policy (everything runs in the browser, progress saves locally, nothing is sent until you hand it to an assistant). Closed: a standalone legal-review FAQ (“Nothing happens without you”) is live and treats ER letters, HR policies, and similar outputs as drafts to verify, with legal or compliance risk flagged on-screen. Closed: affiliate terms were stated in writing by the vendor on 2026-09-13 — 40% flat one-time commission ($39.60 per $99 product), no recurring commission because there is no subscription, 30-day cookie, attribution and payout through Gumroad. Approved and wired 2026-09-13: site ref https://sonaopus.com/?ref=58448483 on Visit and official-pricing CTAs; per-product Gumroad affiliate URLs on the five profession products. Editorial score, cons, and methodology are unchanged. Unresolved: No service address or city/country is published for the seller or data controller; the vendor declined to publish one beyond the Nigeria governing-law clause. The operator name is disclosed.; No data-export feature exists for locally stored workflow data; the Privacy Policy states this and directs users to contact support.; A stale third-party mirror (devadex.com) continues to advertise a retired $59 pre-launch price for the HR product. The mirrored Gumroad account now 404s; the mirror is outside the vendor's direct control. |
| Speechify Audio & Video | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing page checked 2026-09-01. Terms of Service state that Speechify Voice Over Studio AI voice overs may be used for commercial purposes. The Terms grant a non-exclusive, non-transferable, revocable license for personal use of the app on one registered account on one device. The Privacy Policy (updated July 22, 2026) states Speechify does not sell user information and offers both free and paid services. An AI Voice API is available under supplementary terms at speechify.com/terms-ai-voice-api. Re-verified 2026-10-01: Free and Premium ($29/month) are unchanged. The page now exposes an annual billing toggle advertised as saving 60%, which the 2026-09-01 record said was absent; the captured view still rendered monthly amounts. Premium access now also bundles voice typing, AI podcasts and the voice AI assistant, and Speechify publishes a public affiliate program at https://speechify.com/affiliates/ advertising 30% of every sale referred with a 30-day cookie. Unresolved: The pricing page presents monthly amounts with a yearly toggle advertised as saving 60%; no single annual total was exposed on the captured view.; Commercial use is permitted for Voice Over Studio output, but the general app license is for personal use on one device; broader commercial deployment may require additional terms.; The Privacy Policy does not explicitly state whether user content (uploaded text) is used to train Speechify's AI models. |
| Spellbook Legal | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official homepage describes Word-native contract review and drafting, intake-to-signed-contract search, and a zero-retention statement that LLM providers do not train on or store customer data. A public trust center lists SOC 2 Type II and other vendor compliance resources. Treat those as vendor claims to verify in contract review. Re-verified 2026-10-01: The official homepage still publishes no self-serve price list; custom per-seat annual subscriptions and the zero-retention statement remain as recorded. Unresolved: Exact per-seat USD pricing is quote-specific and was not listed on the official homepage.; Zero-retention and training statements should be confirmed in the current DPA/contract, not only marketing copy.; HIPAA/SOC/GDPR resources on the trust center are vendor-published, not independently certified here. |
| Suno Audio & Video | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing checked 2026-08-28. Subscription credits do not carry over; purchased top-up credits do not expire but require an active subscription. Commercial-rights limits depend on plan and terms. Re-verified 2026-10-01: the official pricing page now publishes dollar amounts for the Free, Pro and Premier plans, which the 2026-08-28 check could not retrieve. Subscription credits still do not carry over; purchased top-up credits do not expire but require an active subscription. Unresolved: Commercial-rights limits depend on plan and on the terms of service; the Free plan states it grants no commercial rights.; Annual prices are shown as 'saves $24' / 'saves $72' versus monthly rather than as a single annual total on the captured view. |
| Synthesia Video | 2026-11-03 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Customer Terms effective July 20, 2026 say the customer owns Customer Data, including Inputs and Outputs, while Synthesia owns Synthesia Content. Customers receive a limited license to Synthesia Content during the subscription and, after termination, to the extent that content was incorporated into generated videos, subject to continuing compliance. The terms license Customer Data to Synthesia only as reasonably necessary to provide, maintain, and improve the services, address security/support issues, comply with law, prevent fraud, or as expressly authorized in writing. The privacy policy generally describes Synthesia as processor of Customer Data under customer instructions. Personal-avatar documentation requires a live consent recording matching the person shown; stock avatars are based on actors who consented. The Acceptable Use Policy prohibits rights violations, deceptive impersonation, and disabling provenance mechanisms. It restricts paid social advertising with Stock Avatars unless Synthesia gives express written consent; Custom Avatar advertising requires appropriate consent from the depicted individual. Stock-avatar use in sensitive endorsement-like contexts, news, politics, or current-events content is also restricted. Acceptable-use source: https://www.synthesia.io/legal/acceptable-use-policy Unresolved: The pricing page is internally inconsistent: its annual cards display Starter at $18/month and Creator at $64/month, while its FAQ states $264/year for Starter and $804/year for Creator. Those annual totals do not equal the displayed monthly equivalents, so no corrected value was inferred.; The reviewed public terms permit use of Customer Data to improve the services but do not clearly state whether that includes general model training, nor do they identify a self-serve training opt-out. A stronger no-training claim could not be verified.; Commercial use is not unrestricted: Stock Avatar advertising and sensitive-context uses can require written consent, an Enterprise plan, or both. |
| TAROTAI.CHAT Creative | 2026-10-21 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-20 (14:22 UTC), category 'Creative' (accurate). Editorial entry written independently. One of four products from the same operator, confirmed in writing 2026-10-20 15:34 UTC; the same reply supplied the inference path (WaveSpeed, google/gemini-3.7-flash) and the session-record retention mechanics. Held on an unnamed contracting party and a GA4 claim that does not match its configuration. Re-verified live 2026-10-21: /privacy-policy (dated 2026-09-27) still reads 'We use self-hosted Umami and Google Analytics 4 ... load by default without requiring click consent', no G- measurement ID appears in any served page, and the shipped config carries an empty gaId. PUBLISHED 2026-10-21 with 'no registered company, trading name, postal address or controller identity' and the GA4/gaId mismatch recorded as stated limitations in cons rather than as a gate. Governing law (Hong Kong) and the provider/model disclosure are credited in pros. Unresolved: None recorded |
| TenantCloud Property Management | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing page checked 2026-08-26. Some older third-party pages still show Starter at $18/month; the official page now lists $25/month monthly. Business is customized above the published $100/month starting point. Unresolved: Exact Business extras above the $100/month starting point are quote-specific.; Unit caps per self-serve plan were not enumerated as a single public table on the pricing page during this check.; Marketing customer-count figures are vendor claims. |
| TheFaceRating Creative | 2026-10-24 Pricing ↗ | Not published Privacy ↗ Rights ↗ | submitted via form by vendor; editorial entry written independently. Received through FormSubmit on 2026-10-24 12:49 UTC from vban@mail.grokbot.com, no prior submission under this name or domain in the repository, the hold ledger or any mail folder. Live pages fetched 2026-10-24: /, /tools, /tools/face-shape-detector, /tools/attractiveness-test, /studio, /privacy, /terms, /sitemap.xml and /robots.txt all 200; /pricing, /privacy-policy, /terms-of-service, /about, /contact, /refund-policy, /sample-report, /refunds and /llms.txt all 404 (the nav's Pricing link is the homepage anchor, and Terms section 7 is the only refund text). Product behaviour verified rather than read: the free scan was run in a logged headless Chrome against the vendor's own sample image (/examples/ai-face-rating-sample.jpg) after instrumenting window.fetch and XMLHttpRequest - the scan produced 478/478 landmarks, 78/100 harmony, a face shape and per-region bands with zero outbound requests, and the following step posted to /api/scan/create a body of sha256 + geometry + celebrity match and no photo field when the consent step was declined. A result page was created at /scan/<uuid> with no account. The paywall there lists the same three one-time prices as the homepage and Terms, and clicking Unlock Starter opened a live Stripe checkout for USD 7.99; the checkout was abandoned and no payment was made. Analytics read from the shipped markup: GA4 G-VGC2RBRR2T preloaded and configured inline with anonymize_ip, allow_google_signals false and ad personalization off, plus an unnamed Cloudflare Web Analytics beacon. Domain registered 2026-09-15 (Spaceship), Cloudflare NS johnathan/athena, and no MX record at all, so every published @thefacerating.com address - privacy@ in the policy and Terms - cannot receive mail; the submission address on a separate domain resolves to AWS inbound SES and is the only working route we could identify. An SMTP RCPT probe against that host returned 550 5.7.1 'IP address blacklisted by recipient' for both the candidate address and a random local part, so the probe says nothing about the mailbox - SES is refusing this machine's IP, not the address; the real send left through Gmail's SMTP. Entity 'Vban LLC' with a Sheridan, Wyoming address and Wyoming governing law is named in the Terms but did not surface in a public company search. No name collision of consequence: 'The Face Rating' is descriptive, the nearest properties are the competing products the site itself names (TheFaceReport, Umax, Qoves, PinkMirror, FaceIQ Labs), and the product is already indexed by third-party directories (ainave.com, aisotools.com). Category submitted as Creative - accurate. Unresolved: None recorded |
| TryBestSpecs Creative | 2026-10-25 Pricing ↗ | Not published Privacy ↗ Rights ↗ | Submitted via form by vendor (an AI assistant on the vendor's behalf); editorial entry written independently. Product behaviour read from the served pages and the shipped client bundle: MediaPipe FaceLandmarker with GPU delegate loaded in the browser from cdn.jsdelivr.net and storage.googleapis.com, an analysis request schema carrying base64Image plus optional biometricData, and a billing/history surface behind the account. Legal pages /privacy-policy and /terms-of-service both exist, both dated 28 April 2026, and both print support@trybestspecs.com; /refund-policy, /refunds, /cookies, /cookie-policy, /acceptable-use, /data-rights and /legal/* all 404. Contact route probed against the domain's own Cloudflare Email Routing MX with a control: 250 2.1.0 for support@trybestspecs.com against 550 5.1.1 "Address does not exist" for a random local part, so the mailbox is live and the domain is not a catch-all. Trackers read from the root layout in the served HTML: googletagmanager AW-18136694057 and clarity.ms tag x12jqf0qok, both lazyOnload, present on /privacy-policy itself, with zero consent/cookie-banner strings anywhere. The Stripe/Creem/PayPal/Plausible/Clarity hits in the admin i18n catalogue were checked in context and are translation keys for an admin settings panel, not this vendor's processor list; the Google Ads and Clarity tags are real injected scripts, quoted separately. RDAP: registered 2026-01-10T10:18:40Z, NS aliza/stanley.ns.cloudflare.com (a pair used by no other held submission), MX Cloudflare Email Routing. No famous project, repo or incumbent product trades as TryBestSpecs or BestSpecs; footprint is a Product Hunt hunt, a make.rs profile and one directory listing. No affiliate or partner page exists (/affiliates, /affiliate, /partners, /referral all 404), so no tracking URL was invented. One served-string trap avoided: a search-engine snippet of /pricing still says "3 free credits upon registration" while the live page, its meta description and its JSON-LD all say 1 credit - the fetch wins and no contradiction was raised with the vendor. Unresolved: None recorded |
| Tunii Audio & Video | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Live re-check 2026-09-18 at tunii.net/pricing and the same figures appear in the page FAQ. The previous record said paid packs had no public pricing page; that was wrong - the packs are published with dollar amounts, and they are one-time purchases rather than a subscription. Terms and privacy pages are live and were linked on this check. Unresolved: None recorded |
| TypingMind AI Agents | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing page (typingmind.com/buy) checked 2026-09-01. TypingMind is a BYOK (bring-your-own-key) LLM frontend; users connect their own OpenAI, Anthropic, Google, or local Ollama API keys and pay providers directly for usage. The app is not open-source; self-hosting is available with a license key but only the compiled code is provided. The homepage states data is stored locally by default and conversations are not used for AI training. The Premium tier was advertised at 50% off ($99 from $198) as a limited-time offer. Unresolved: The $99 Premium price is a limited-time 50% off discount; the regular price is $198 and may revert.; API costs are separate and depend on the user's chosen model provider and usage; TypingMind does not publish or control those rates.; The self-hosted version provides compiled code only, not source code; modification rights are not granted. |
| UMEVO Meetings | 2026-10-27 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via the site form by the vendor; editorial entry written independently from live pages on the site's Pacific working date 2026-10-27. Held since 2026-10-25 and re-verified after the vendor's 2026-10-27 reply. Hardware prices read from the store's own products.json, not from the submission, which quoted Note Plus at $119 against the store's $109. The AI plan figures were re-read on both product pages the same day; they disagree with each other (see unresolved_claims) and both are recorded rather than one being picked. The two legal sets are recorded separately because they disagree: /pages/privacy-protection-policy section 10 names the third-party companion app AI DVR Link, while /policies/privacy-policy omits it and carries a 1560 Broadway address the vendor says is out of date; /pages/terms-of-service names UME Technology Inc. while /policies/terms-of-service names Huasengdi Technology Co., Ltd. of Sheridan WY. The app policy at record.szaicy.com was fetched and read on the vendor's link: Dongguan Kinghal Electronic Technology Co., Ltd., US storage, about one month of retention after account cancellation, Sales@kinghal.com for deletion. No training-use statement and no consent/recording-law warning appear on any umevo.ai page checked. Colorado entity 20251272831 verified in good standing. Not hands-on tested. Unresolved: Which paid ladder applies after the included AI period: Note ONE quotes Pro $7.99/mo or $69.99/yr (1,800 min) and MAX $19.99/mo or $149.99/yr (unlimited); the Note Plus page quotes 'Pro $2.50/mo or Max $5.83/mo, billed annually', both as unlimited. Minute packs likewise differ roughly 4x: $2.49/$9.49/$44.99/$79.99 on Note ONE against $0.59/$2.59/$10.99/$19.99 on Note Plus. The vendor was asked which is authoritative on 2026-10-27.; Which entity is the contractual counterparty: the footer terms name UME Technology Inc., the store policy terms name Huasengdi Technology Co., Ltd. (Sheridan, WY) as the entity accountable under the EULA, and both governing-law clauses name no state. The vendor stated in writing (2026-10-27) that UME Technology Inc. of Denver is the seller and operator and Colorado law governs, and that the older 1560 Broadway notice will be aligned. Not yet reflected on the pages.; Whether a retention period and a training-use statement will be published on umevo.ai itself: the vendor confirmed ~1 month after account cancellation (via the third-party Kinghal policy) and 'no training' in writing, but no umevo.ai page states either, and the Kinghal policy is linked from nowhere on the site.; Whether the Note Plus page's 'Pro $2.50/mo / Max $5.83/mo' and its pack prices are stale copy from an earlier plan or a genuine second ladder. |
| Udio Audio & Video | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official pricing checked 2026-08-28. Paid plans add editing, audio upload, stem download, and higher concurrent-generation limits. Re-verified 2026-10-01: the monthly credit allocations changed from the 2026-08-28 record. Standard now lists 2,400 monthly credits (recorded as 1,200) and Pro lists 6,000 (recorded as 4,800). Plan prices and top-up prices are unchanged. Unresolved: None recorded |
| Whop CLI Automation | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Not published | Independently verified 2026-08-27 from official Whop docs (docs.whop.com/developer/cli) and Whop blog (whop.com/blog/cli). CLI launched August 2026. Blueprints launched same period. Re-verified 2026-10-01: The Whop site still carries no separate CLI subscription; the record's 'free CLI install, standard Whop platform fees' remains accurate. Unresolved: None recorded |
| Words into Music Audio & Video | 2026-10-21 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Submitted via form by the vendor 2026-10-21 21:38 UTC; editorial entry written independently from live pages on 2026-10-21. Pricing read from /pricing/ and cross-checked against the site's own llms.txt, which states the same tiers, annual totals and pack prices. /privacy/ (updated 18 October 2026) and /terms/ (updated 15 October 2026) read in full; the privacy policy names Cloudflare, Convex, Better Auth, Google, IPLocate, Replicate, Stripe, Resend, Migadu and Crisp with their roles, states retention, and lists the Umami event fields. Contact route verified: domain MX is Migadu, contact@wordsintomusic.com returned 250 in an SMTP RCPT probe while a random local part returned 550. Parent brand savuko.com lists the product and names Poland in its own app terms, which the product's terms do not. The '4.7/5 from 241 customer ratings' figure in the footer and JSON-LD is not repeated in this listing because no public source was found. Unresolved: None recorded |
| You.com Research | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official materials document Smart Agent, premium agents, multiple third-party models, custom agents, file handling, and search/research APIs. The privacy policy says prompts and user-created/shared content sent to third-party AI providers are not used to develop, improve, or train those third-party models. Separately, You.com's terms permit You.com to use Content to provide, maintain, train, and improve its own services and develop products. The terms assign You.com's rights in outputs to the user, subject to applicable law and excluding third-party output. Official API documentation expressly permits commercial applications when the terms are followed and API keys and user data are secured. Re-verified 2026-10-01: The official pricing page still publishes the recorded API rates (Web Search $5/1,000 calls, Contents $1/1,000 pages, Answer $5/1,000 calls, Research from $12/1,000 calls) and no consumer chat-plan prices; no drift detected. Unresolved: Current consumer chat-plan names and prices were not published on the reviewed public pricing page.; The output assignment does not cover third-party output, whose separate terms may apply.; Commercial-use confirmation located in official support expressly concerns the API, not every consumer-chat use case. |
| Zapier AI Automation | 2026-09-18 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | AI by Zapier supports built-in models or a connected customer AI-provider account, structured output fields, knowledge sources, and app actions as tools. Zapier's April 9, 2026 Terms define Inputs and Outputs as Customer Content and state that customers retain ownership of Customer Content. Zapier receives a limited license needed to provide, maintain, operate, improve, and support the service. Zapier may derive de-identified data sets from Customer Content and use them to improve products, including through model training. Customers may submit the linked opt-out form; Company and Enterprise accounts are automatically opted out under the current Terms. Usage Information may be used to develop and improve Zapier products, but the current clause does not explicitly say that Usage Information is used for model training. Unresolved: The official AI by Zapier help page is internally inconsistent about Free-plan access: its availability panel marks Free unavailable, while a later section discusses the Standard tier on a Free plan.; A current official Agents Pro dollar price was not reliably exposed in the retrieved pricing-page text.; Zapier's Terms do not promise that AI output is unique, copyrightable, or non-infringing.; Third-party model handling depends on Zapier's AI Functionality Subprocessors and, for bring-your-own-model configurations, the selected provider's terms. |
| n8n Automation | 2026-10-01 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official documentation says n8n workflows can connect providers including OpenAI, Anthropic, and Google, add tools and memory, and combine models. n8n's AI Terms, last updated March 20, 2026, say n8n will not use Customer Data to train n8n-operated AI models, retain it beyond immediate AI processing, or share it except as needed to deliver customer-configured AI features. Usage Data may be used to improve services only when it does not contain Customer Data or identify the customer or users. Customer-configured third-party model providers receive data under their own terms. As between n8n and the customer, the customer owns AI Output to the extent permitted by law; n8n receives a limited operational license and will not use AI Output to train n8n-operated AI models. Re-verified 2026-10-01: the published cloud allowances changed from the 2026-08-25 record. Starter now lists 1,600 Assistant credits/month (recorded as 2,300 AI credits) and Pro lists 4,000 or up to 9,600 (recorded as 5,700 or up to 13,700). Free trials are listed as 800 Assistant credits plus 100 AI credits. Execution counts and euro plan prices are unchanged. Unresolved: The no-training commitment is expressly limited to AI models operated by n8n; third-party LLM treatment depends on the provider selected by the customer.; The official terms qualify output ownership with 'to the extent permitted by applicable law' and do not promise uniqueness, copyrightability, or non-infringement.; The retrieved pricing page distinguishes Assistant credits from model-provider usage; customer-built workflows calling external LLMs may incur separate provider charges.; Assistant and Agents features are labelled preview on the official pricing page and their allowances can change. |
| v0 Development | 2026-10-25 Pricing ↗ | Docs ↗ Privacy ↗ Rights ↗ | Official documentation describes v0 as an AI agent for producing full-stack apps and agents, with Git integration, a full editor, previews, backend connections, and Vercel deployment. v0 may use user prompts or generated content with third-party models and learning systems to improve products. Enterprise content and content from customers who opt out in Vercel settings are not used for training. Vercel platform-service customer data or code is not used to train, improve, or fine-tune v0 models. The FAQ says Vercel does not own code generated from user prompts, while warning that output may resemble other output or third-party IP and must be evaluated by the user. Pricing terminology on the retrieved current documentation uses 'Plus'; some indexed official snippets still use the former name 'Team'. Re-checked 2026-10-25: new Free accounts receive a one-time $5 signup credit rather than a recurring $5 monthly credit, and Premium at $20/month remains closed to new signups. Business at $100 per user per month includes the same $30 monthly credit allowance as Plus. Unresolved: The official page says model token rates can be viewed in the model selector or pricing page, but the extracted live documentation did not consistently expose a complete current rate table.; The FAQ does not promise copyrightability, uniqueness, non-infringement, or an IP indemnity for generated code.; The effect of deleting content already incorporated into a trained model is limited: the FAQ says existing models will not be updated after such a deletion request. |
Methodology: source links are stored in data/tool_sources.json and regenerated into this public ledger. Editorial ratings are independent; affiliate relationships never determine scores. Prices and policies change — confirm the vendor source before purchasing.