Privacy Policy

Effective date: October 1, 2026

This policy explains how track by track GmbH handles personal data when you use the llms.txt Generator, either on www.llms-txt-generator.de or through its MCP server, for example as the "llms.txt Generator" plugin in ChatGPT. A German version is available at /datenschutz; it additionally covers website-only features in more detail.

1. Who is responsible

track by track GmbH, Schliemannstr. 23, 10437 Berlin, Germany
Email: hi@geo-tool.com · Phone: +49 30 21923869
Company details: Imprint

2. Summary

No account is needed. We do not create user profiles and do not store generated results.
The ChatGPT plugin receives only what its tools need: a domain or URL, or the text of an llms.txt file.
We never request your chat history, name, email address or other account data from ChatGPT.
We do not sell personal data, do not use it for advertising and do not send it to language-model providers.

3. Personal data we collect

3.1 When you use the ChatGPT plugin or another MCP client

Tool inputs: the website (domain or URL) you ask about, or the llms.txt text you provide, and optionally a hostname. These usually do not identify you, but a domain or text can relate to you, for example a personal website.
Client metadata: ChatGPT sends technical metadata with each tool call, including an anonymized user identifier assigned by OpenAI (openai/subject) and, where provided, locale or session information. We use only the anonymized identifier, as described in section 4.
Connection data: IP address of the calling server, time of the request, requested endpoint, user agent and response status, processed by our hosting provider.

3.2 When you use the website

Generator inputs and connection data as above.
Usage statistics: page views and generator events (for example a successful crawl with the checked hostname, page count and validation score). Full URLs, paths and crawled content are not included.
Contact form: your name, email address and message, only if you submit the form and consent.
Optional monitoring: email address, domain and technical snapshots, only after you subscribe and confirm by double opt-in.

3.3 Website content we read

To produce a result, the service reads publicly accessible pages of the domain you specify: HTML, titles, meta descriptions, headings, links, canonical tags, robots.txt, sitemaps and the published llms.txt. Through the plugin it honors robots.txt, identifies itself with the user agent llms-txt-generator.de, never logs in and never bypasses access controls.

We do not intentionally collect payment data, health data, government identifiers or passwords.

4. Why we use it

To provide the requested result (generate, validate or check an llms.txt). Legal basis: performance of the requested service, Art. 6(1)(b) GDPR.
To prevent abuse and keep the service available: we turn the anonymized ChatGPT user identifier, or otherwise the IP address, into a truncated SHA-256 key and count tool calls per key. Legal basis: legitimate interest, Art. 6(1)(f) GDPR.
To improve the website through the usage statistics in 3.2. These are not collected for plugin calls. Legal basis: legitimate interest, Art. 6(1)(f) GDPR.
To answer your enquiries and run the monitoring you subscribed to. Legal basis: consent, Art. 6(1)(a) GDPR, which you can withdraw at any time.

5. Who receives it

The client you use. Results go back to the application that called the tool. In ChatGPT, OpenAI processes them as part of your conversation under the OpenAI Privacy Policy; OpenAI is responsible for that processing.
Hosting: Vercel Inc., 340 S Lemon Ave #4133, Walnut, CA 91789, USA, as our processor. Transfers to the USA are covered by Vercel's Data Processing Addendum.
Website analytics: a self-operated analytics endpoint at analytics.polymarkt.de (website only).
Contact form: Slack, where enquiries arrive in our workspace (website only).
Monitoring: Upstash (storage, primary region Frankfurt) and Brevo (email delivery), only for subscribers.
The website you check sees ordinary requests from our crawler, as with any web visit.

We do not share personal data with language-model providers, data brokers or advertisers.

6. How long we keep it

Tool inputs and results: not stored. The application does not log submitted URLs, llms.txt text or crawled content; data exists only for the duration of the request.
Rate-limit key: held in server memory only and deleted after ten minutes.
Hosting logs: kept according to Vercel's log retention for the configured plan.
Contact enquiries: deleted once handled, unless legal retention duties apply.
Monitoring: unconfirmed sign-ups are removed after 48 hours; subscription data and snapshots are deleted immediately when you unsubscribe.

7. Your choices and rights

You decide which website or text you submit. Do not include personal information in llms.txt text you want checked.
In ChatGPT you can disconnect or disable the plugin at any time in your ChatGPT settings.
You can block the website analytics script in your browser.
You may request access, correction, deletion, restriction or portability of your data, object to processing based on legitimate interest and withdraw consent. Write to hi@geo-tool.com.
You may lodge a complaint with a data protection supervisory authority, for example the Berlin Commissioner for Data Protection and Freedom of Information.

8. Security

All connections use HTTPS. The crawler rejects private, internal and local network addresses, limits response sizes and treats page content as untrusted data. Rate limits protect the service from misuse.

9. Children

The service is a technical tool for website owners and is not directed at children.

10. Changes

We update this policy when data flows or providers change and show the effective date at the top.