Skip to main content
All requests to the CUDO Compute API must be authenticated with an API key in the Authorization header unless explicitly documented otherwise.

Overview

The CUDO Compute API uses a simple API key based bearer scheme: Header format:
Your API key uniquely identifies you and inherits the permissions of your user or project context. Keep keys secret and never embed them in client-side code or public repositories.

Create an API key

You can create API keys in two ways:
  1. Via the CUDO Compute console (recommended for most users)
  2. Programmatically with the POST /v1/api-keys endpoint
Successful response (200):
The key field is only returned once at creation time. Store it securely now; you cannot retrieve it later.

Using your API key

Include the header on every request:
Python example:

Error handling

Common authentication-related HTTP status codes:
Our errors follow the Google AIP 193 specification. Please see the error codes and formats in the Errors section of the documentation.

Best practices

Rotation workflow example

  1. Create a new key.
  2. Deploy the new key (update environment variables / secret stores).
  3. Verify all services function with the new key.
  4. Revoke the old key.
This minimizes downtime and risk.

Environment variables

Store your key locally in a shell profile or an .env file:
Then reference it in tooling and scripts instead of hardcoding.

Rate limiting & retries

If you receive HTTP 429, implement exponential backoff (e.g., wait 1s, 2s, 4s, 8s…) and respect any Retry-After header (if present). Do not continuously retry invalid keys-fix the credential instead.

Troubleshooting

Currently the API uses a single global bearer scheme named bearerAuth with header Authorization. There is no separate refresh or OAuth flow.

Next steps

You’re authenticated-explore the API reference or jump into building clusters, machines, and more.