Is a free AI API safe to use? You don’t have to trust us

Every fact you can verify is right here; if you still have doubts, run it in a way that doesn’t require trusting us at all.

Your concerns, and what we do about them

Your concerns, and what we do about them
Will my code be kept or sold?Normal requests leave no content behind — only things like time and usage are recordedUse only your own computer
Can the owner of the computer doing my work see my content?Every frame is encrypted; that computer doesn’t know who you areUse official nodes only
Could a free API make Claude Code run rogue commands or steal files?The API only returns text; commands are run by Claude Code on your computer, which asks first by defaultRun it in a container
What if my key leaks?Keys can only call models and can’t reach your servers or accountRevoke with one click; unused guest keys expire automatically
Could I be charged unexpectedly?Guests have no payment method; a balance is used only when points run shortSet a daily limit per app

How the encryption works

  1. Your browser to TutuCompute: HTTPS.
  2. TutuCompute to the computer doing your work: an extra encryption layer inside the channel, with every frame encrypted on its own.
  3. Every connection uses fresh temporary keys, so recorded traffic can’t be decrypted later.
  4. The computer verifies the platform with a built-in public key, so no one in the middle can swap in their own key.

Run Claude Code / Codex in a container

Start a container that mounts only the current project, then paste and run the command copied from the console.

docker run --rm -it \
  -v "$PWD":/work -w /work \
  --cap-drop ALL --security-opt no-new-privileges \
  node:22 bash
  • It can’t see your home folder, SSH keys or cloud credentials
  • Everything is gone when you exit
  • Commit with git first so you can always roll back

FAQ

Related