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

# Audalis MCP

> Ask Claude, ChatGPT, or another AI assistant about your calls, bookings, contacts, and follow-ups in plain language.

Audalis MCP is a hosted [Model Context Protocol](https://modelcontextprotocol.io) server. Connect it to an AI assistant, and you can ask questions about your Audalis workspace in plain language, such as "Which callers asked for service this week but didn't book?"

Audalis hosts the server and handles sign-in. There's nothing to install and no API key to manage.

**Server URL:** `https://api.audalis.io/mcp`

<Note>
  **The connector is read-only.** It can't send messages, place calls, book or change appointments, or edit anything in Audalis. A bad prompt can read your data, but it can't act on your customers.
</Note>

## What you can ask about

<CardGroup cols={2}>
  <Card title="Calls" icon="phone">
    Calls answered by your team or by Arlo, with caller, outcome, sentiment, and a summary. Filter to missed calls only.
  </Card>

  <Card title="Bookings" icon="calendar-check">
    Appointments Arlo booked in Jobber or Square, plus callers who asked for service but didn't end up booked.
  </Card>

  <Card title="Contacts" icon="address-book">
    Find a customer by name, company, email, or phone number, and see their summary, recent calls, and open tasks.
  </Card>

  <Card title="Follow-ups" icon="list-check">
    Open follow-up tasks, soonest due first, including the ones Arlo created from calls.
  </Card>
</CardGroup>

See [Tools](/help/developers/mcp/tools) for every tool, its parameters, and its limits.

## Things to try

Once connected, ask your assistant things like:

* "What calls did we miss yesterday?"
* "Which callers asked for service this week but didn't book? Who should I call back first?"
* "Catch me up on Maria Lopez before I call her back."
* "What follow-ups are due before Friday?"
* "Summarize how Arlo handled calls this week. Anything it struggled with?"
* "How many jobs did Arlo book in Jobber last month?"

The assistant chooses which tools to call and can combine several to answer one question.

## How it works

<Steps>
  <Step title="Add Audalis to your AI assistant">
    Paste the server URL into Claude, ChatGPT, Claude Code, Cursor, or another assistant that supports remote MCP servers. See [Connect to Audalis MCP](/help/developers/mcp/connect).
  </Step>

  <Step title="Sign in and allow access">
    Your browser opens Audalis. Sign in, confirm the workspace and app name, and click **Allow**. Your password is entered only on Audalis and is never shared with the assistant.
  </Step>

  <Step title="Ask in plain language">
    The assistant reads from the workspace you approved. Every request uses your Audalis account, and access stops right away if you disconnect or lose your owner or admin role.
  </Step>
</Steps>

## Who can connect

* You need an active Audalis workspace.
* Only workspace **owners** and **admins** can connect an app, because a connection can read the whole workspace's call history.
* If connecting fails with "Connected apps are not enabled for this workspace", email [support@audalis.co](mailto:support@audalis.co) and we'll turn it on.

## Next steps

<CardGroup cols={3}>
  <Card title="Connect" icon="plug" href="/help/developers/mcp/connect">
    Set up Claude, ChatGPT, or another client.
  </Card>

  <Card title="Tools" icon="wrench" href="/help/developers/mcp/tools">
    Every tool, parameter, and limit.
  </Card>

  <Card title="Security" icon="shield-check" href="/help/developers/mcp/security">
    Keep customer data safe.
  </Card>
</CardGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.