Build with Auro.

Use Auro through an OpenAI-compatible API. Every signed-in account gets 100,000 free tokens across input and output.

Quick start

Set your key in an environment variable and send a message. The current model is auro-v5; GET /v1/models lists its name.

Limits that keep Auro available

Free allowance

100,000 input + output tokens per account, shared across all its keys. This is a one-time allowance. Replacing a key keeps your usage.

One at a time

One request per account at a time, including requests waiting in the queue. Use n: 1.

Request rate

Up to 6 requests per minute per account.

Request size

Up to 4,096 input tokens, including system prompts and tools, and 1,024 output tokens. Free access accepts text.

Shared GPU

The API runs at most 2 generations at once across all keys and shares the chat’s queue. Chat visitors have higher GPU priority than free API requests.

Token accounting

Input is counted before generation. Output is capped to your remaining balance. If a started request disconnects or fails without final usage, its reserved token budget is counted.

Requests and responses

Send Authorization: Bearer YOUR_KEY and Content-Type: application/json to POST /v1/chat/completions. Supply model and a messages array with role and content. With no system message, Auro uses its own prompt.

Common options include temperature, top_p, max_tokens, stop, and stream. Streaming returns server-sent events ending in [DONE]; set stream_options: {"include_usage": true} to receive usage.

{
  "choices": [{"message": {"role": "assistant", "content": "Hello!"},
               "finish_reason": "stop"}],
  "usage": {"prompt_tokens": 18, "completion_tokens": 3, "total_tokens": 21}
}

Illustrative response. Your token count includes any default system prompt.

When a request is refused

400 · Input

Check the message format, token limits, and n: 1.

401 · Key

Supply an active API key from your signed-in account.

402 · Allowance

Your remaining tokens cannot cover this request. Shorten the prompt or contact Alan on X for more access.

429 · Busy

Wait for your current request to finish, or follow the Retry-After header before retrying.

503 · Unavailable

Auro or token accounting is temporarily unavailable. Try again shortly.