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

# Quivly MCP Server

> Connect Claude, ChatGPT, Cursor, and other AI tools to your Quivly customer data with one URL: https://mcp.quivly.ai/connect. OAuth sign-in, read-only, org-scoped.

Quivly ships a hosted [MCP](https://modelcontextprotocol.io) server, so any MCP-capable AI tool — Claude, ChatGPT, Cursor, VS Code, and others — can query your customer data directly. Ask Claude "which of my accounts are most at risk right now?" and it answers from your live Quivly data.

**Server URL:**

```
https://mcp.quivly.ai/connect
```

The first connection opens your browser once to sign in to Quivly (OAuth). Answers are read-only and scoped to your organization — you only ever see your own org's data.

In-app setup guides for each client live at **Settings → AI → MCP**.

## Client setup

<Tabs>
  <Tab title="Claude">
    **Settings → Connectors → Add custom connector**, paste the URL, click Add, and complete the one-time sign-in. On Team/Enterprise plans, an Owner adds it under Organization settings → Connectors.
  </Tab>

  <Tab title="Claude Code">
    ```bash theme={null}
    claude mcp add --transport http quivly https://mcp.quivly.ai/connect
    ```

    Then run `/mcp` to sign in. Add `--scope user` to use it across all projects.
  </Tab>

  <Tab title="Cursor">
    Add to `mcp.json`:

    ```json theme={null}
    { "mcpServers": { "quivly": { "url": "https://mcp.quivly.ai/connect" } } }
    ```
  </Tab>

  <Tab title="VS Code">
    Add to `.vscode/mcp.json`:

    ```json theme={null}
    { "servers": { "quivly": { "type": "http", "url": "https://mcp.quivly.ai/connect" } } }
    ```
  </Tab>

  <Tab title="ChatGPT">
    Enable Developer mode, then **Settings → Connectors → Create**, name it Quivly, and paste the URL.
  </Tab>

  <Tab title="Other clients">
    Add Quivly as a remote (Streamable HTTP) MCP server with the URL above and complete the OAuth sign-in. Windsurf, Zed, and Cline configs are in the in-app setup guides.
  </Tab>
</Tabs>

## What it can do

Around 40 tools cover customer profiles, health scores, revenue, usage, contracts, calls and transcripts, support tickets, opportunities, conversations, notes, notebooks, projects, cross-customer aggregation, plus live Slack, Salesforce, and HubSpot lookups. One tool writes: `create_notebook` generates an [AI notebook](/product/notebooks) using one of your [skills](/product/skills). Everything else is read-only.

See the full [tool reference](/developers/mcp-tools).

## Usage analytics

**Settings → AI → MCP → Usage** shows who's connecting and what they ask: users, sessions, breakdowns by client app and by integration, filters, and an "Ask your usage" chat for querying it in natural language. Per-user breakdowns are admin-only.

## FAQ

<AccordionGroup>
  <Accordion title="Is it safe to connect an AI tool to our customer data?">
    Access requires each user's own Quivly sign-in, answers are scoped to their organization, everything except notebook creation is read-only, and every tool call is logged in the usage analytics.
  </Accordion>

  <Accordion title="Does each teammate need to connect separately?">
    Yes — the OAuth sign-in is per user, so access and audit trails follow the individual.
  </Accordion>

  <Accordion title="Can the AI modify or delete our data?">
    No. The only write operation is creating a new notebook. Nothing can be edited or deleted through MCP.
  </Accordion>
</AccordionGroup>