> For the complete documentation index, see [llms.txt](https://boundaryai.gitbook.io/boundaryai-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://boundaryai.gitbook.io/boundaryai-docs/api-and-webhooks/mcp.md).

# Connect AI assistants (MCP)

BAI Analytics exposes a read-only **MCP server** (Model Context Protocol), so AI assistants can answer questions like *"what are the top complaints in Customer Support EU this month?"* directly from your analysed feedback. MCP is the open standard AI tools use to reach external data; Claude, and a growing set of clients, speak it natively.

```
https://app.boundary-ai.com/api/mcp
```

Authentication is an API key with the **`mcp_read`** scope (or `all`), passed as a bearer token. `mcp_read` exists so you can hand an assistant a key that can query analytics but can **never push, create, or send**. The other scopes deliberately don't include it, so an existing integration key doesn't silently become an LLM query surface.

***

### What the assistant can do

Six read-only tools:

| Tool                  | Answers                                                         |
| --------------------- | --------------------------------------------------------------- |
| `list_sources`        | What feedback groups and sources exist, and how big they are.   |
| `get_source_analysis` | A source's sentiment distribution, themes, and monitor matches. |
| `search_feedback`     | Find actual comments matching a query.                          |
| `list_themes`         | The themes across a source or group.                            |
| `list_monitors`       | Custom Monitoring monitors and their match counts.              |
| `get_sentiment_trend` | How sentiment moved over time.                                  |

Everything is scoped to the key's organisation, and it's read-only end to end: nothing an assistant does through MCP can modify your data.

***

### Setup

{% tabs %}
{% tab title="Claude Code" %}

```bash
claude mcp add bai-analytics --transport http \
  https://app.boundary-ai.com/api/mcp \
  --header "Authorization: Bearer inpk_live_..."
```

{% endtab %}

{% tab title="Any MCP client" %}
Configure a **streamable HTTP** server:

```json
{
  "url": "https://app.boundary-ai.com/api/mcp",
  "headers": {
    "Authorization": "Bearer inpk_live_..."
  }
}
```

{% endtab %}
{% endtabs %}

Then just ask: *"Using BAI Analytics, what changed in customer sentiment since last month, and which monitor is firing most?"*

{% hint style="info" %}
Clients that only support OAuth-based remote connectors (rather than a bearer header) aren't supported yet; OAuth support is planned. Any client that can send an Authorization header works today.
{% endhint %}

***

### Good practices

* **Create a dedicated `mcp_read` key per assistant** and label it accordingly, so you can revoke one assistant's access without touching your integrations.
* **Treat the assistant's access like a team member's read access.** It can see everything the organisation's analytics contain.
* Rate limits apply as on any key; see [Authentication & API keys](/boundaryai-docs/api-and-webhooks/authentication.md).