> For the complete documentation index, see [llms.txt](https://docs.brainnetwork.app/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.brainnetwork.app/participate/use-the-api.md).

# Use the API

Change one URL. Run AI on Brain.

## Key

Create an API key under [Account → API keys](https://brainnetwork.app/account). Keys look like `brain_sk_…`, are shown once, and are stored as sha256.

## Request

{% tabs %}
{% tab title="curl" %}

```bash
curl https://brainnetwork.app/v1/chat/completions \
  -H "Authorization: Bearer $BRAIN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen2.5-1.5b-instruct",
    "stream": true,
    "messages": [{ "role": "user", "content": "explain entropy in one line" }]
  }'
```

{% endtab %}

{% tab title="Python" %}

```python
from openai import OpenAI

client = OpenAI(base_url="https://brainnetwork.app/v1", api_key="brain_sk_...")
r = client.chat.completions.create(
    model="brain/auto",
    messages=[{"role": "user", "content": "hi"}],
)
print(r.choices[0].message.content)
print(r.model_extra["brain"]["target"], r.model_extra["brain"]["receiptId"])
```

{% endtab %}

{% tab title="JavaScript" %}

```js
import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://brainnetwork.app/v1", apiKey: process.env.BRAIN_API_KEY });
const r = await client.chat.completions.create({
  model: "brain/auto",
  messages: [{ role: "user", content: "hi" }],
});
console.log(r.choices[0].message.content);
```

{% endtab %}
{% endtabs %}

## Choosing a model

* A **node model id** from [/models](https://brainnetwork.app/models), such as `qwen/qwen2.5-1.5b-instruct`, runs on Brain Nodes and nowhere else. The page shows live node counts and measured speed per model.
* **`brain/auto`** lets BRAIN AUTO choose across browser compute, Brain Nodes, operator cloud and external models by the published score.

Optional fields:

| Field        | Values                                                           |
| ------------ | ---------------------------------------------------------------- |
| `mode`       | `AUTO` (default) · `CHEAP` · `FAST` · `QUALITY` · `BROWSER_ONLY` |
| `privacy`    | `PUBLIC` · `STANDARD` (default) · `PRIVATE`                      |
| `maxCost`    | USD ceiling; excludes classes whose known cost exceeds it        |
| `maxLatency` | ms ceiling; excludes classes whose known latency exceeds it      |

## Privacy

Node models run on hardware BRAIN does not operate, so they are `privacy: public`. A `STANDARD` or `PRIVATE` request pinned to a node model returns `400 privacy_conflict` instead of being routed somewhere else. With `brain/auto`, privacy is a hard filter applied before scoring.

## What comes back

Standard OpenAI-shaped chunks or a completion object, plus:

**Headers:** `brain-request-id`, `brain-target`, `brain-latency`, `x-brain-receipt`, and for node work `brain-node-id`, `brain-region`.

**The `brain` object** (and the stream's closing `event: brain`): route, model, cost, receipt id, and for node work `brain.node = { id, region, reliability, computeClass, routing, verification }`. `routing` is the sentence explaining why that node won. `verification` is the shadow re-run result if one happened.

## When nothing can run it

`503` with every target's exclusion reason, for example:

```json
{
  "error": "no_capacity",
  "targets": {
    "NATIVE_NETWORK": "0 of 4 nodes can serve qwen/qwen2.5-7b-instruct right now (N-895D3F34: VRAM 8 GB < 20 GB required; N-CF66A0F8: offline; …)",
    "EXTERNAL_MODEL": "not configured",
    "BROWSER_NETWORK": "capability chat not supported"
  }
}
```

BRAIN never answers from a target it did not select.

## Verifying a receipt

Fetch the public key from [`/api/coordinator/signer`](https://brainnetwork.app/api/coordinator/signer). Rebuild the canonical body (`v, jobId, nodeId, model, inputTokens, outputTokens, executionMs, timestamp, requestHash, responseHash, hardwareClass, cost`) with sorted keys, sha256 it, and check the ed25519 signature. `verifyReceipt()` in the repository does exactly this and nothing more.

## Limits

Per-IP fixed-window rate limits on every public route; plan rate limits per key (10 requests a minute on Free). Prompt size limits are enforced by the gateway; image parts are rejected rather than dropped. Job output is capped at 200,000 characters.

Full reference: [/developers](https://brainnetwork.app/developers).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.brainnetwork.app/participate/use-the-api.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `build a script that syncs our docs to a CMS` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.