# Developer docs

The Continuum API runs leading open AI models in Australian data centres. It works with the OpenAI SDKs and any tool built for them, so most code runs after changing two settings: the address and the key.

```text
Base URL   https://api.continuumlabs.ai/v1
Auth       Authorization: Bearer cl-your-key
```

## How it works

You send a request to our gateway in Sydney. The gateway checks your key and your credit, passes the request to the model, and streams the answer back. It counts the tokens and charges your account. What you send, and what the model replies, is not stored.

## Start here

1. [Get a key](/console), then tell us what you're building and we'll add trial credit.
2. Follow the [quickstart](/docs/quickstart) to make your first request.
3. Copy a working agent from the [cookbook](https://github.com/continuumlabs-ai/cookbook): contract review, research with citations, invoice processing and more.

## What the API supports

| Feature | Supported |
|---|---|
| Chat completions, `POST /v1/chat/completions` | Yes |
| Streaming | Yes |
| Tool calling | Yes, on models that support it |
| JSON output | Yes, on models that support it |
| List models, `GET /v1/models` | Yes |
| Embeddings, `POST /v1/embeddings` | Yes |
| [Allowances](/docs/allowances), [receipts](/docs/receipts) and [statements](/docs/statements) | Yes |
| [Management API](/docs/management-api), `/v1/workspaces`, `/v1/keys`, `/v1/statements` | Yes |
| Images and audio endpoints | Not yet |
| Batch endpoint | Not yet |
| Fine-tuning | Through [Continuum Lab](/lab), not the API |

> Using an AI coding assistant? Point it at [https://www.continuumlabs.ai/llms.txt](https://www.continuumlabs.ai/llms.txt), or use **Copy page as Markdown** at the top of any page, and it can write working code against this API.