> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fingerprint.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Automation Intelligence API

> The Automation Intelligence API (in public preview) gives you the tools to determine whether traffic is legitimate and should be accepted by your application.

<Warning>
  **Feature in Public Preview**

  This feature is currently in a Public Preview testing phase. All feedback is welcome! If you encounter any issues, please [contact](https://fingerprint.com/support/) our support team.
</Warning>

The API detects automation tools like AI Agents, AI Assistants, and AI Browsers. Additionally, it provides IP intelligence like geolocation and data center detection.

Automation Intelligence is derived from HTTP request metadata that reaches your server. It does not require the use of a Javascript agent to collect browser context. The API is fast, with average response times of less than 30ms, making it a great fit for edge, pre-origin or middleware contexts. The API is platform-agnostic and can be used with different CDN providers, cloud platforms, or any server backend. If you are a Cloudflare user, consider using [Flow](/docs/flow-deployments#using-a-flow-worker-to-detect-bots-at-the-edge).

## Regions

The Automation Intelligence API is available in the **Global**, **EU** and **Asia (Mumbai)** regions:

| Region        | Base URL                 | Server Location    |
| ------------- | ------------------------ | ------------------ |
| Global        | `https://api.fpjs.io`    | Global             |
| EU            | `https://eu.api.fpjs.io` | Frankfurt, Germany |
| Asia (Mumbai) | `https://ap.api.fpjs.io` | Mumbai, India      |

Use the **Base URL** dropdown in the right column of the endpoint to select the correct base URL according to your workspace region.

## Authentication

Use an `Authorization: Bearer` header to authenticate to the API. All unauthenticated requests will return a `HTTP 403 Forbidden` response.

### Authorization: Bearer HTTP header

When making an API request, add the `Authorization: Bearer SECRET_API_KEY` HTTP header with your [secret API key](https://dashboard.fingerprint.com/api-keys?type=api).

## Forwarding request headers

Forward **all** of the original request's headers to the API, preserving their original order and capitalization. Detection accuracy depends on receiving the request as the client sent it.

<Warning>
  **Send authentication and session headers with empty values — do not drop them**

  Headers that carry authentication or session data, such as `Authorization`, `Cookie`, and `Set-Cookie`, must still be included, but with their value set to an empty string. Omitting them entirely changes the shape of the request and can affect detection. Never forward the real secret values.
</Warning>

For example, given an incoming request with these headers:

```http Incoming request theme={"theme":"github-dark-dimmed"}
User-Agent: openai-mcp/1.0.0
Authorization: Bearer sk-live-abc123
Cookie: session=9f8a7b6c
Accept: application/json
```

forward them like this, blanking only the sensitive values while keeping every header present and in order:

```json Forwarded headers theme={"theme":"github-dark-dimmed"}
{
  "headers": [
    { "name": "User-Agent", "value": "openai-mcp/1.0.0" },
    { "name": "Authorization", "value": "" },
    { "name": "Cookie", "value": "" },
    { "name": "Accept", "value": "application/json" }
  ]
}
```

## Rate limiting

If you exceed the rate limit when making an API call, you'll get an HTTP 429 error:

```json JSON theme={"theme":"github-dark-dimmed"}
{
  "error": {
    "code": "too_many_requests",
    "message": "too many requests"
  }
}
```

Retry the request after a short interval. The response may include a `Retry-After` header indicating how long to wait.

## Trying it out

You can try calling the Automation Intelligence API directly from this reference:

1. You are going to need a Secret API Key. You can create one in your Fingerprint **Dashboard** > [**API Keys**](https://dashboard.fingerprint.com/api-keys).
2. Select the [Collect Intelligence](/reference/server-api-edge) endpoint.
3. Click **Try it**.
4. Set **Authentication** to your secret API key.
5. Enter a **URL**, some example **Headers**, an HTTP **Method**, and an **IP** address.
6. Make sure the **Base URL** corresponds to your workspace region.
7. Click **Send**!

A real API response will appear in the **Response** section. Alternatively, you can view the prepared response examples there.

You can also point an API client like Postman to the [OpenAPI Schema](https://fingerprintjs.github.io/fingerprint-pro-server-api-openapi/schemas/fingerprint-server-api-v4-with-examples.yaml) directly.

## Related guides

<Columns cols={3}>
  <Card title="Bot Detection Overview" icon="bot" href="/docs/bot-detection/overview">
    How Fingerprint detects AI agents, browsers, and automation tools, and what each result means.
  </Card>

  <Card title="Bot Directory" icon="book-open" href="/docs/bot-detection/bot-directory">
    The bots and AI agents Fingerprint can detect, with their result types.
  </Card>

  <Card title="Web Bot Auth Implementation" icon="shield" href="/docs/bot-detection/web-bot-auth-implementation">
    How to implement Web Bot Auth signing so your agent is verified as a signed agent.
  </Card>
</Columns>