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

# AI Coding Agents

> Wire Dashtray into Claude Code, VS Code, Zed, Windsurf, Gemini CLI, and any MCP client

# AI Coding Agents

Once the Dashtray MCP server is configured, any AI coding agent can text
your phone when it finishes a task. Kick off a long refactor, close the
laptop, and get the ping:

> "Implement the export feature, and when you're done notify me on mobile
> through Dashtray."

Setup is **one-time and global** - the server runs on your machine via
`npx`, so the same config works in every project you open afterwards.

## Prerequisites

1. A Dashtray account (no plan needed to sign up; the API requires an
   active subscription - see [pricing](/guides/pricing))
2. An API key from [Settings → API keys](https://web.dashtray.app/settings/keys)
   * copy it immediately - keys can't be viewed again

## Claude Code (terminal)

The simplest setup of any client - one command, scoped to your user so it
works in every project:

```bash theme={null}
claude mcp add dashtray --scope user \
  --env DASHTRAY_API_KEY=sk_YOUR_KEY_HERE \
  -- npx -y @dashtray/mcp-server
```

Verify with `claude mcp list` (shows `dashtray - ✓ connected`), then just
ask: *"Ping my phone when the tests pass."*

## VS Code (Copilot)

Create `.vscode/mcp.json` in your project (or add it to your user
`mcp.json` for every project):

```json theme={null}
{
  "servers": {
    "dashtray": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@dashtray/mcp-server"],
      "env": {
        "DASHTRAY_API_KEY": "sk_YOUR_KEY_HERE"
      }
    }
  }
}
```

<Note>
  VS Code uses `"servers"` (not `"mcpServers"` like the other clients)
  and wants an explicit `"type": "stdio"`. Easy to mix up when copying
  from another tool's config.
</Note>

## Zed

Add to `~/.config/zed/settings.json`:

```json theme={null}
{
  "context_servers": {
    "dashtray": {
      "source": {
        "id": "dashtray",
        "command": "npx",
        "args": ["-y", "@dashtray/mcp-server"],
        "env": {
          "DASHTRAY_API_KEY": "sk_YOUR_KEY_HERE"
        }
      }
    }
  }
}
```

## Windsurf

Add to `~/.codeium/windsurf/mcp_config.json`:

```json theme={null}
{
  "mcpServers": {
    "dashtray": {
      "command": "npx",
      "args": ["-y", "@dashtray/mcp-server"],
      "env": {
        "DASHTRAY_API_KEY": "sk_YOUR_KEY_HERE"
      }
    }
  }
}
```

## Gemini CLI

Add to `~/.gemini/settings.json`:

```json theme={null}
{
  "mcpServers": {
    "dashtray": {
      "command": "npx",
      "args": ["-y", "@dashtray/mcp-server"],
      "env": {
        "DASHTRAY_API_KEY": "sk_YOUR_KEY_HERE"
      }
    }
  }
}
```

## Any other MCP client

If your tool speaks MCP over stdio, the server config is always the same
shape: run `npx -y @dashtray/mcp-server` with `DASHTRAY_API_KEY` in the
environment. Claude Desktop, Cursor, and Cline have dedicated guides:

<CardGroup>
  <Card title="Claude Desktop" icon="brain" href="/integration/claude-desktop">
    claude\_desktop\_config.json setup
  </Card>

  <Card title="Cursor" icon="terminal" href="/integration/cursor">
    Cursor MCP setup
  </Card>

  <Card title="Cline" icon="code" href="/integration/cline">
    Cline setup
  </Card>
</CardGroup>

## Troubleshooting

**Server fails to start inside a monorepo** (`sh: 1: dashtray-mcp: not
found` or a silent startup failure): npm's runner can resolve the `npx`
binary against your workspace's own `node_modules` instead of the fetched
package. Two fixes:

1. Install the server globally and point your client at it directly:

   ```bash theme={null}
   npm install -g @dashtray/mcp-server
   ```

   then replace the command with `"command": "dashtray-mcp"` (no args).

2. Or set the server's working directory to a neutral location (for
   example your home directory) in your client's MCP config, if it
   supports a `cwd` field.

**"Invalid API key"**: the key is wrong or revoked - generate a new one
in [API keys](https://web.dashtray.app/settings/keys).

**Notifications not arriving**: check quiet hours in the app's
notification settings, and note the API rate limits (Starter: 100
requests/min, Pro: 1,000 requests/min).


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