Privacy Policy

Last updated: 2026-05-28

1. Overview

AI FactScan is designed to help users inspect citation source quality in AI-generated answers. The extension is local-first and minimizes data transmission.

AI FactScan does not sell personal data, does not collect browsing history, and does not use submitted data for advertising.

2. Local processing

When you paste text or select an AI answer block, AI FactScan processes the content in your browser to extract citation URLs, DOI links, arXiv links, and PubMed links.

General AI answers and full selected page text are not sent to the FactScan API for source-quality lookup.

3. Data that may be stored locally

AI FactScan may store the following on your device:

• extension settings
• language preference
• supporter license status
• usage count for free daily scans
• a randomly generated device ID (factscan_device_id), used to verify device binding for Supporter access
• cached source-quality database files
• cached public benchmark/source-signal files
• bundled CRED-1 negative-signal data

4. Citation checks and external requests

AI FactScan may make limited external requests when needed:

• DOI resolver checks to doi.org, only when the citation itself is a doi.org/... link
• arXiv checks to arxiv.org/abs/..., only when the citation itself is an arXiv link
• PubMed checks to pubmed.ncbi.nlm.nih.gov, only when the citation itself is a PubMed page
• public benchmark refreshes, such as Vectara Hallucination Leaderboard data
• optional FactScan API domain lookup, which sends citation domains only

General article URLs are classified by source domain. AI FactScan does not scan general URL paths for hidden DOI-like strings in order to mark them invalid.

5. FactScan API lookup

When online source lookup is enabled, AI FactScan may send citation domains and a device identifier to the FactScan API. For example, domains such as imf.org, theguardian.com, or openai.com may be sent.

The device identifier is a randomly generated ID stored locally on your device. It is used only to verify that the license is being used within the permitted device limit. It is not linked to your identity, browsing history, or any personal information.

The API lookup does not send the full AI answer, full selected page text, or browsing history.

To maintain service quality, the FactScan API may track per-license daily domain lookup counts. This data is stored in a private database and is not shared with third parties.

6. Feedback submissions

Feedback is sent only when you click the feedback button and submit the feedback form.

A feedback submission may include:

• citation domain
• citation URL
• current grade
• suggested grade
• feedback type
• suggested classification
• evidence type
• evidence URL
• optional note written by you
• page host
• locale

Feedback is used for review and source-quality database improvement. Feedback does not automatically change public source ratings.

7. Feedback infrastructure and notifications

Feedback may be stored in a private review database and may generate a private Discord notification for the AI FactScan operator. The notification contains submitted feedback fields and a review ID.

Discord is used as an internal notification channel, not as a public rating source.

To reduce spam and abuse, the FactScan API may apply rate limiting and derive a short-lived hashed submitter key from request metadata such as IP address and User-Agent. Raw IP addresses are not stored in the feedback table.

8. Payment and supporter access

If you purchase or support AI FactScan through a third-party payment provider, payment information is handled by that provider. AI FactScan does not directly store full payment card details.

AI FactScan may store supporter license status or license keys needed to unlock supporter features.

9. Third-party services and references

AI FactScan may interact with third-party services only for limited purposes described above, such as citation resolver checks, public benchmark refreshes, feedback notification, payment processing, and optional FactScan API lookup.

The bundled source database references the Vectara Hallucination Leaderboard, the CRED-1 dataset (licensed under CC BY 4.0), and domain signals derived from the content-farm-list project. These are external, independently maintained resources. AI FactScan does not share any user data with them.

Third-party services may have their own privacy policies and terms.

10. Children's privacy

AI FactScan is not directed at children under the age of 13. We do not knowingly collect any information from children.

11. Changes to this policy

If this policy changes materially, we will update the date at the top of this page. Continued use of the extension after changes constitutes acceptance of the updated policy.

12. Contact

For privacy questions, data questions, or support requests, contact: addfocus247@gmail.com