Privacy information
This describes the current software and account flow. Account sign-in is available, but beta enrollment is closed. Final publisher details, retention periods and a monitored privacy contact remain under review before broad beta access.
Your work and chosen providers
Auton keeps code and local journals on your machine except the prompts and context sent to the AI providers and services you configure. Those services process requests under their own terms. Local work does not mean an AI task stays entirely offline.
Account and billing information
Auton's account service processes identity and profile information from GitHub or Google, linked sign-in providers, sessions and device authorization requests. It stores beta enrollment, access deadlines, suspension decisions and policy audit records to determine access.
Payments are disabled and beta access requires no payment details. Future paid Lite and Pro billing is planned through Paddle. Before purchases are enabled, this page will describe the payment information, customer and subscription identifiers, and reconciliation records used for that service.
The account service uses request addresses for rate limiting when served through its configured trusted proxy. Administrative access changes are recorded in an audit log.
This website and browser demo
Optional website reporting is off by default. You can allow or turn it off using Privacy choices in the footer. We remember only that choice in this browser's local storage. Do Not Track and Global Privacy Control keep reporting off, even if you previously allowed it. Revoking consent stops new reports, clears the memory queue and cancels pending requests; it cannot withdraw a report already received.
What optional website reports contain
If you allow reporting, our self-hosted Umami service receives counts for a small list of public pages and clicks that open pricing, download information or sign-in information. These clicks do not prove an installation, account or purchase. Reports omit actual URLs, query strings, fragments, referrers, page text, form values, screen size, language and account identifiers.
Our self-hosted GlitchTip service receives a generic error category, website release and known bundled JavaScript filenames with line and column numbers. We discard raw error messages, unknown or external script locations, local variables, request bodies, console activity, user identity and page contents. We do not record sessions, screenshots, AI prompts or responses, source code, or terminal content.
Requests necessarily pass through Cloudflare and our hosting infrastructure, which receive network addresses and browser request headers. Umami may derive broad device, location and rotating visitor/session statistics from these headers; we do not send a distinct visitor or account ID. Disabling optional reporting does not disable hosting security logs. Reports are limited in the browser and failures never prevent using the site. Error retention is configured for 30 days; analytics retention and backup deletion are still being verified before broad beta access.
The browser demo uses sample folders and simulated activity; it does not access repositories or contact AI providers. Fonts and the reporting code are served locally, and no advertising tracker is included.
Questions and requests
A verified privacy contact and instructions for account data requests will be published before public account access. See the current support status.