Begins w/ AI API

What's new: GPT 5.2 and Gemini 3 Flash is now available

• Quickstart

  Make your first request in under 2 minutes

• API Reference

  Full endpoint documentation

• Supported Models

  Available models and how to choose

• Examples

  Complete working code (Python + JavaScript)

---

What is this?

This is an async AI API. You send a request, we queue it, you poll for the result.

There's no streaming. No webhooks. Just HTTP requests and a polling loop.

How it works

sequenceDiagram
    participant Client
    participant API as BWA API
    participant Queue as Processing Queue

    Client->>API: POST /v1/ai<br/>{model, prompt}
    API->>Queue: Queue request
    API-->>Client: 202 Accepted<br/>{request_id}

    Note over Client: Wait & Poll

    loop Poll until complete
        Client->>API: GET /v1/ai?req_id=xxx
        alt Still processing
            API-->>Client: {status: "queued"}
        else Complete
            API-->>Client: {status: "completed", result}
        end
    end

    Note over Client: Use result

The flow is dead simple:

1. POST your request to /v1/ai with a model and prompt
2. Get back a request_id immediately (HTTP 202)
3. Poll /v1/ai?req_id=<request_id> until status changes from queued to completed
4. Retrieve your result from the response