Privacy Policy

Last updated 28 June 2026.

W1 is a VS Code / Cursor extension that runs an AI coding agent on a hosted backend (app.w1lab.com). Some data necessarily leaves your machine; this explains what, why, and how to control it.

What we collect

Model inference. Every model call is proxied through the W1 backend — that is how W1 runs without provider keys on your machine. Your prompt and the code context the agent reads or edits are sent to the backend and on to the model provider for that call.

Run traces — on by default, opt-out. When a run finishes, W1 uploads a trimmed trace (your prompt, the model's plan, tool and command observations, and a code diff capped at 20,000 characters) to diagnose runs and improve the agent, attributed to your account via your sign-in token. No provider keys or secrets are included. The full-fidelity trace stays local in .w1/runs/. Turn upload off with w1.shareTraces: false.

Usage analytics. Pseudonymous product telemetry (run started/completed, token counts, latency, model, errors) under a hashed per-user id. No source code or file content is sent.

What we do not collect

• Provider API keys or secrets — there are none on your machine to send.
• Your source code beyond the inference context and the ≤20k-character diff in a trace.
• File contents in usage analytics.

Your controls

• w1.shareTraces: false — stops run-trace upload (traces stay only on your machine).
• VS Code telemetry.telemetryLevel: off — disables both trace upload and usage analytics.

Data retention

Uploaded traces are stored for 30 days, then expire automatically.

Contact

Questions or deletion requests: updates@updates.w1lab.com.