Privacy Policy
For the Fact-Check — YouTube, articles & podcasts browser extension
(Chrome Web Store item apknlmiimjjchbeppbgepfnmkhghafil), and for the
claimpit.com website. Last updated 5 August 2026.
Fact-Check is a browser extension that checks the claims in the YouTube video, web article, or podcast episode you are viewing, using an AI model you configure. This policy explains what data is collected, how it is handled and used, where it is stored and for how long, and who it is shared with.
There is no account and no sign-up. The extension acts only when you click it, it stores your settings and API key on your own device, and the extension itself contains no analytics, ads, or tracking — it never reports your browsing to us.
Please read this before you use the extension. Fact-Check is a public fact-checking project, not a private tool. When you fact-check a video, article, or podcast episode, the result is published on claimpit.com as a public web page — the source's title, link, thumbnail, and every claim and verdict, together with the transcript excerpts that back them. These pages are listed in our sitemap, so search engines can index them.
Only fact-check public content. Do not use the extension on private, internal, paywalled, or otherwise sensitive pages, on anything behind a login, or on anything you would not want to appear on a public website.
1. What data is collected
Content you choose to fact-check
This is collected only for the specific page you act on, and only at the moment you click. The extension does not read pages in the background, and it does not collect anything from pages you never fact-check.
| When you check a… | The extension collects |
|---|---|
| YouTube video | the video id and watch URL, the video title, and the video's time-coded transcript. If the video has no captions, its audio is transcribed for us (see section 4). |
| Web article | the page URL, the page title, the main article text, and the URL of the article's lead image. |
| Podcast episode | the episode page URL, the show's RSS feed URL, the episode title, show notes/description, artwork URL, episode length, the audio file URL, and the episode audio itself, which is transcribed to text. |
Your settings and API key
You bring your own API key for your chosen AI provider. The extension stores, on your own device only: your chosen provider, your API key for each provider you configure, the model and reasoning level you picked, and your grounding and claim-limit preferences.
What is never collected
- No accounts, names, email addresses, or other personal identifiers. The extension has no sign-up and no user ID.
- No analytics, telemetry, or advertising in the extension, and no browsing-history collection. It never tells us which pages you visit, and it never reports a page you did not click Fact-check on.
- Never your API key. It is sent only to the AI provider you chose, to authorize your own requests. We do not receive, see, or store it.
- No keystrokes, form fields, passwords, cookies, financial or health information, or location data.
2. How the data is handled and used
The data above is used for exactly one purpose — running and publishing the fact-check you asked for — in these steps:
- Transcription. For a video without captions, or a podcast episode, the audio is converted to text so its claims can be checked. Nothing else is done with the audio.
- Checking. The text is sent to the AI provider you selected, which extracts individual claims and judges each one against live web sources.
- Publishing. The resulting claims and verdicts, and the source details listed above, are written to our database and published on claimpit.com. There they are also indexed into our public claim, topic, and public-figure pages — so a claim can be listed under the person who said it and the topic it belongs to.
- Caching. The stored result is reused when anyone checks the same video, article, or episode again, so the check does not have to be paid for and run twice.
- Your settings and API key are used solely to make the request to your chosen provider on your behalf.
We do not use any of this data to build a profile of you, to train our own models, for advertising or marketing, or to determine creditworthiness or for lending purposes. We do not use it for anything unrelated to fact-checking the source you asked about.
3. Where data is stored, and for how long
-
On your device. Your provider choice, API keys, model and reasoning
settings, and a local cache of results are stored in your browser's extension storage
(
chrome.storage.local). They stay there until you clear the extension's storage or uninstall it, which deletes all of it — including your API keys. - On our servers. The source details, transcript, claims, and verdicts are stored in a Cloudflare D1 database we operate, hosted on Cloudflare's network. This is the same database that serves the public claimpit.com pages.
- Retention: indefinite. Published fact-checks are kept for as long as the site exists — that is the point of the project, and there is no automatic expiry. They are deleted when we remove them, including on request; see section 6.
- Audio and article text are held only transiently while a check runs; what is retained afterwards is the resulting transcript text, source details, and verdicts, not the original audio file.
4. Who the data is shared with
Published publicly
As described in the notice above, the fact-check itself — the source's title, link, thumbnail or artwork, its claims and verdicts, and the transcript excerpts quoted as evidence — is published on claimpit.com and visible to anyone, including search engines. Nothing about you is published: results carry no user identifier, because we never collect one.
Service providers who process content on our behalf
Each of these receives only what it needs to do its step, and each is governed by its own privacy policy:
| Who | Receives | Why |
|---|---|---|
| OpenAI, Google (Gemini API), or Anthropic — whichever you select | the article text or transcript, and your API key | to extract and judge the claims. Only the one provider you configured is contacted. |
| Cloudflare | everything we store, plus podcast episode audio | hosting for our database and Workers, and Workers AI (the @cf/openai/whisper-large-v3-turbo speech model), which transcribes podcast audio. |
| Supadata | the YouTube video id | to retrieve a video's captions, or generate a transcript when it has none. |
| Our own transcription server | audio of caption-less YouTube videos | a self-hosted speech-to-text service we run ourselves. It is not a third party and passes the audio to no one. |
| Wikidata / Wikimedia | the names of public figures mentioned in claims | to identify the people a published claim is attributed to. No user data is sent. |
What we never do
- We do not sell your data, and we do not transfer it to third parties for advertising, marketing, or any purpose beyond the fact-checking use described here.
- We do not transfer it to determine creditworthiness or for lending purposes.
- We disclose data to no one else, except where we are legally required to.
5. Analytics on this website
This section is about the claimpit.com website only — not the extension, which contains no analytics whatsoever. We count visits so we can tell which pages are worth keeping, using Plausible Analytics, an EU-based, open-source alternative to Google Analytics chosen specifically because it does not identify or follow people:
- No cookies, and no use of browser cache or local storage — which is why this site has no cookie banner.
- No persistent identifier, and no profile. Your IP address is never stored. It is combined with your browser's user-agent string and a secret salt to produce a one-way hash that counts you as one visitor for the day. That salt is deleted and regenerated every 24 hours, so you cannot be recognized the next day, and never across other websites.
- What is measured: the page viewed, the site that referred you, and coarse technical details — browser, operating system, device type, and approximate location no finer than city, derived from the IP address. Nothing else.
- Where it goes: the data is processed and stored in the EU, by Plausible.
These requests come from claimpit.com, not from a third-party domain. The analytics script and the measurement requests are served from this site and forwarded to Plausible by our server. We do this so the measurements still work when a content blocker is active, and you should know two things about it: your IP address is passed on to Plausible in that forwarding step, where it is hashed and discarded as described above; and because the requests are first-party, a content blocker will not stop them.
Opting out. If you would rather not be counted, open your browser's
developer console on this site and run
localStorage.plausible_ignore = "true". That flag lives on your own
device, and while it is set this site will not send any measurement at all.
6. Your choices, and deleting data
- Nothing happens unless you click. The extension collects a page's content only when you press Fact-check on it. Not using it on a page means nothing about that page is collected.
- Local data: clear the extension's storage or uninstall it, and your settings, API keys, and local cache are gone.
- Published data: email us to have a fact-check removed from claimpit.com and from our database. Tell us which source, and we will delete it. You do not have to explain why, and there is no account to prove.
- Access and other rights: because we store no identifier for you, we cannot look up "your" data — but you can ask us anything about the data we hold for a given source, and we will answer.
7. Permissions the extension requests, and why
- storage — save your provider choice, API key, model settings, and locally cached verdicts on your device.
- activeTab and scripting — add the fact-check UI and read the article, transcript, or podcast episode, only on the tab where you clicked.
- Host access to your AI provider's API (OpenAI, Google, Anthropic), to
www.youtube.com(add the button and read the transcript), and to our cache and transcription service on*.workers.dev.
8. Children
Fact-Check is not directed at children, and we do not knowingly collect data from them. There is no account, so we never ask for an age or any personal detail.
9. Changes to this policy
If what we collect, how we use it, where we store it, or who we share it with changes, we update this page and change the "last updated" date above. Material changes to how the extension handles your data will also be reflected in the extension itself.
10. Contact
Questions, deletion requests, or anything else about this policy: hello@claimpit.com. We are the developer of the Fact-Check extension and the operator of claimpit.com, and we answer these ourselves.