Privacy policy

Information we handle

We handle account and sign-in details, the prompts and project material you submit to the service, agent activity, model usage, and support messages. For purchases, we receive payment status, order references and billing details needed to provide access and resolve issues. Cashfree handles checkout and payment credentials; please do not send card details to us by email.

Why we use it

We use this information to authenticate you, run requested coding tasks, provide model access, measure usage, prevent abuse, process purchases, restore access when a billing or usage error occurs, answer support requests, and understand whether our coding tools complete users' tasks. For event participants, we may compare sign-ups, project progress, model usage and agent reliability to improve future events. The terminal and editor send bounded activity metadata such as tool names and outcomes, token counts, duration and file-change counts; this diagnostic record does not include prompt text, source code, file paths, tool arguments or tool output. We do not collect full local conversation transcripts for event analytics unless you separately choose to share them. Managed inference still sends the relevant prompt and context to the selected model provider to answer your request.

Service providers

Account, hosting, model inference and payment providers process information needed for their parts of the service. Prompts and relevant project context sent to a managed model may be processed by the selected model provider. If you bring your own model API key, your chosen provider's terms and privacy practices also apply. The location of processing depends on the provider and deployment; do not assume all processing happens in India.

Retention and your choices

We keep records needed to operate accounts, maintain billing and usage history, resolve disputes and meet applicable obligations. You can ask us about your information, request correction or deletion where applicable, and raise a privacy concern by emailing ibrahim@code.in. We may need to retain some records where law or a legitimate operational obligation requires it.

Website and app analytics

We use Google Analytics to understand visits and a few successful sign-in, installation-request and checkout-start actions. Page addresses use fixed route templates; project identifiers, query strings, fragments, private page titles and referrers are excluded. Event fields are limited to the website or app surface, English or Arabic, and fixed sign-in method, installation channel or checkout-kind labels. Google receives standard browser/device information and uses first-party analytics cookies. Its requests expose your IP address to Google; this is not anonymous traffic. Advertising signals and advertising personalization are disabled. The Google tag runs in a separate fixed Code.in document rather than the account app or checkout document.

Turn website analytics off below or in the app’s Settings. This browser preference applies across code.in and app.code.in and is saved in a preference cookie and, when available, browser storage. We respect Do Not Track and you can also use a browser analytics blocker.

Terminal and editor analytics

The terminal and editor send Google a bounded operational summary for an agent turn: client surface and version, platform, terminal outcome, observed duration, model-call and token counts, tool call/result/failure counts, dropped-event counts and, when measured, changed-file counts and time to first visible text. Managed model names come from our server catalog; custom provider names are replaced with a generic label. A random installation identifier links these analytics summaries; a derived numeric analytics session value does not expose your local session identifier. These Google analytics records exclude account IDs, email addresses, conversation/run IDs, prompts, code, paths, diagnostic events and tool names, arguments or output. Turn this off with /settings analytics off in the terminal or “Share anonymous usage” in the editor’s General settings. Do Not Track and the editor’s telemetry preference also apply. Private service diagnostics and requested managed inference remain separate from Google analytics.