# Agent-Native Analytics MCP server

> Lets an AI client connect analytics data sources and generate charts and dashboards from natural-language prompts.

- Repository: https://github.com/BuilderIO/agent-native
- Page: https://hostedmcp.org/servers/builderio/agent-native
- Official: yes
- Transport: remote
- Stars: 7696
- License: MIT
- Latest version: 1.0.0

## Install

### Claude Code

```sh
claude mcp add --transport http agent-native https://analytics.agent-native.com/mcp
```

### Cursor (`~/.cursor/mcp.json`)

```json
{
  "mcpServers": {
    "agent-native": {
      "url": "https://analytics.agent-native.com/mcp"
    }
  }
}
```

### VS Code (`.vscode/mcp.json`)

```json
{
  "servers": {
    "agent-native": {
      "type": "http",
      "url": "https://analytics.agent-native.com/mcp"
    }
  }
}
```

### Codex (`~/.codex/config.toml`)

```toml
[mcp_servers.agent-native]
url = "https://analytics.agent-native.com/mcp"
```

### Gemini CLI (`~/.gemini/settings.json`)

```json
{
  "mcpServers": {
    "agent-native": {
      "httpUrl": "https://analytics.agent-native.com/mcp"
    }
  }
}
```

## FAQ

### What can Agent-Native Analytics do with my data?

It lets you ask questions of your data and build dashboards, connecting sources like Amplitude and Mixpanel.

### How do I create a new agent app with Agent-Native?

Run the create command with npx and pick a template, for example the chat template.

### What database backends does Agent-Native support?

It uses PostgreSQL in production and PGlite for local development.

### How are capabilities shared between the agent and the UI?

Each capability is defined once as a shared action, so the agent uses it as a tool and the UI calls it from code with the same validation and permissions.