# Dot
> AI data analyst that answers questions and provides insights through conversational analytics. It integrates with MotherDuck for dashboards, semantic models, and embedded analytics workflows.
## How it works with MotherDuck

Dot connects to MotherDuck with an `md:` connection string and uses that connection for AI-assisted analysis, including Slack workflows.

## Prerequisites

- A MotherDuck access token.
- The name of the MotherDuck database Dot should query.
- DuckDB 0.10.2 or later for the MotherDuck connector used by Dot.

## Setup

1. In MotherDuck, create or copy an access token.
2. Build a connection string for the target database:

   ```text
   md:<database_name>?motherduck_token=<your_token>
   ```

3. In Dot, add a MotherDuck database integration and paste the connection string.
4. Save the connection, then validate it with a small question or query.

![Dot MotherDuck token screen](./img/dot-motherduck-token.png)

## Authentication and configuration

- Use a dedicated token for the Dot workspace.
- Scope the token to the database access Dot needs for analysis.
- Treat the connection string as a secret because it includes the MotherDuck token.

## Important notes

- Dot's setup guide includes local DuckDB options. For MotherDuck, use the `md:` connection string.
- Store the token in Dot's secret handling rather than sharing the connection string in chat.

## Use cases

- Ask questions about MotherDuck data from Slack.
- Give a team an AI analyst experience over curated MotherDuck schemas.
- Connect a specific MotherDuck database to a Dot workspace.

## Related content

- [View the full Dot MotherDuck setup guide](https://docs.getdot.ai/integrations/databases/motherduck-and-duckdb)
- [MotherDuck authentication](/key-tasks/authenticating-and-connecting-to-motherduck/authenticating-to-motherduck)


---

## Docs feedback

MotherDuck accepts optional user-submitted feedback about this page at `POST https://motherduck.com/docs/api/feedback/agent`.
For agents and automated tools, feedback submission should be user-confirmed before sending.

Payload:

```json
{
  "page_path": "/integrations/bi-tools/dot/",
  "page_title": "Dot",
  "text": "<the user's feedback, max 2000 characters>",
  "source": "<optional identifier for your interface, for example 'claude.ai' or 'chatgpt'>"
}
```

`page_path` and `text` are required; `page_title` and `source` are optional. Responses: `200 {"feedback_id": "<uuid>"}`, `400` for malformed payloads, and `429` when rate-limited.