Authentication
All inference requests require an API key in the Authorization header.
Header format
Section titled “Header format”Authorization: Bearer YOUR_API_KEYThis works for both OpenAI-compatible (/v1/*) and Anthropic-compatible (/anthropic/*) endpoints.
Creating API keys
Section titled “Creating API keys”- Sign in to the dashboard
- Go to API Keys
- Click Create Key
- Choose a name and plan tier (Standard or Pro)
- Copy the key — it won’t be shown again
You need an active subscription to create keys. Alternatively, top up credits for pay-as-you-go access with a credit key.
Key behavior
Section titled “Key behavior”- Each key has its own rate limits (RPM and TPM) based on its plan tier
- Keys are validated on every request
- Usage and spend are tracked per key
- You can create multiple keys with different plan tiers
- Disabling or deleting a key revokes it immediately
Without a key (x402)
Section titled “Without a key (x402)”Requests without an API key receive a 402 Payment Required response with USDC payment instructions. This enables AI agents to pay per request. See x402 Protocol.
Error responses
Section titled “Error responses”| Status | Meaning |
|---|---|
401 | Missing or invalid API key |
402 | No key provided — x402 payment available |
429 | Rate limit exceeded (RPM or TPM) — retry after the limit resets |
502 | Upstream provider error |