Register your app to claim a project and a token, point your LLM client at the gateway, then manage your app’s purpose and watch its verdicts and spend here, no console account needed. Everything on this page is also available via the API.
One call creates your project and mints your token (shown once).
Already registered? Enter your project id and token.
Both are kept in this tab only (sessionStorage) and sent as the x-awacs-token header. Lost your token? Recovery is admin-assisted today.
Store the token now. It is shown exactly once and only its hash is kept server-side.
Your project is registered. Choose your provider, load the configuration into your app, then use the generated client for its model calls.
Environment previews use a token placeholder. The environment download includes your current AWACS token: store it as a secret. Keep your provider key in your existing secret settings. Load these values before using the client snippet.
The connection check sends a synthetic inspection request to AWACS. It does not call your model provider. Send a request from your application afterward to confirm its gateway integration.
Connection has not been checked in this session.
Foundry Anthropic needs a separate operator-approved binding; use its dedicated integration instructions.
| Model | Calls | Tokens in / out | Est. cost |
|---|
| Month | Est. spend | vs prev | Calls | Tokens in / out |
|---|
| Age | Direction | Verdict | Model | Category | Latency |
|---|
| Name | Kind | Hint | Created | Last used | Status |
|---|