Privacy

AI Token Saver fetches a public model guide. Its tools accept only a platform label, an instruction mode, or a previously seen public Revision. They do not accept prompts, documents, student records, credentials, or account settings.

The application does not store user profiles, conversations or submitted text, and does not intentionally log request bodies. It caches the validated public guide for up to 15 minutes. Hosting providers may process network metadata such as IP addresses, request paths and timestamps according to their own policies. ChatGPT or Codex also processes your chat under your account's terms and workspace rules.

The service fetches guide data from GitHub Pages, raw GitHub or jsDelivr. It does not send your prompt to those sources. No inference service is called by this plugin.

For support or a privacy question, open an issue without sensitive data in the public repository.