{
  "markdown": "# Fivetran MCP Server\n\n<!-- mcp-name: io.github.YimingYAN/fivetran-mcp -->\n\n[![PyPI version](https://img.shields.io/pypi/v/fivetran-mcp.svg)](https://pypi.org/project/fivetran-mcp/)\n[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)\n[![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)\n[![Built by Claude Code](https://img.shields.io/badge/Built%20by-Claude%20Code-blueviolet)](https://claude.ai/code)\n\n> **Note:** This repository is built and maintained entirely by [Claude Code](https://claude.ai/code), Anthropic's AI coding assistant.\n\nAn [MCP (Model Context Protocol)](https://modelcontextprotocol.io/) server for controlling [Fivetran](https://fivetran.com/) data pipelines. Enables AI assistants like Claude to manage syncs, monitor connection status, and control data pipelines through natural language.\n\n## Features\n\n- **List connections** - View all Fivetran connections with status\n- **Check sync status** - Get detailed status for any connection including tasks and warnings\n- **Trigger syncs** - Start syncs on demand\n- **Historical resync** - Trigger full data resync or resync specific tables\n- **Pause/Resume** - Control connection scheduling\n- **List groups** - View all destination groups\n- **Test connection** - Diagnose connectivity and configuration issues\n- **Schema visibility** - View schemas, tables, and columns for any connection\n- **Reload schema** - Refresh schema configuration from the source\n\n## Quick Start\n\n### Step 1: Get Fivetran API Credentials\n\n1. Log in to [Fivetran Dashboard](https://fivetran.com/dashboard)\n2. Click your **username** (top right corner)\n3. Click **API Key**\n4. Click **Generate API key**\n5. Copy both the **API Key** and **API Secret** (secret shown only once!)\n\nSee [Fivetran API Getting Started](https://fivetran.com/docs/rest-api/getting-started) for more details.\n\n### Step 2: Store Credentials\n\nAdd to your `~/.env.local` (or equivalent):\n\n```bash\n# Fivetran API\nexport FIVETRAN_API_KEY=\"your-api-key\"\nexport FIVETRAN_API_SECRET=\"your-api-secret\"\n```\n\nThen reload:\n```bash\nsource ~/.env.local\n```\n\n### Step 3: Verify Credentials\n\nTest that your credentials work:\n\n```bash\ncurl -s -X GET \"https://api.fivetran.com/v1/account/info\" \\\n  -H \"Accept: application/json\" \\\n  -H \"Authorization: Basic $(echo -n \"$FIVETRAN_API_KEY:$FIVETRAN_API_SECRET\" | base64)\"\n```\n\nExpected response:\n```json\n{\"code\":\"Success\",\"data\":{\"account_id\":\"...\",\"account_name\":\"...\"}}\n```\n\n### Step 4: Configure Claude Code\n\nAdd to your `~/.claude.json`:\n\n```json\n{\n  \"mcpServers\": {\n    \"fivetran\": {\n      \"type\": \"stdio\",\n      \"command\": \"uvx\",\n      \"args\": [\"fivetran-mcp@latest\"],\n      \"env\": {\n        \"FIVETRAN_API_KEY\": \"${FIVETRAN_API_KEY}\",\n        \"FIVETRAN_API_SECRET\": \"${FIVETRAN_API_SECRET}\"\n      }\n    }\n  }\n}\n```\n\n### Step 5: Restart Claude Code\n\nRestart Claude Code to load the new MCP server. You should now have access to Fivetran tools.\n\n## Available Tools\n\n### Connection Management\n\n| Tool | Description |\n|------|-------------|\n| `list_connections` | List all connections, optionally filtered by group |\n| `get_connection_status` | Get detailed status including tasks, warnings, and sync details |\n| `trigger_sync` | Start a sync for a connection (optional `force` flag) |\n| `trigger_resync` | Trigger full historical resync |\n| `resync_tables` | Resync specific tables only (e.g., `[\"schema.table_name\"]`) |\n| `pause_connection` | Pause a connection |\n| `resume_connection` | Resume a paused connection |\n| `list_groups` | List all groups/destinations |\n| `test_connection` | Run diagnostic tests to identify connectivity/configuration issues |\n\n### Schema & Table Visibility\n\n| Tool | Description |\n|------|-------------|\n| `get_connection_schema` | Get schema info with optional table filter (includes columns when filtering) |\n| `get_schema` | Get complete raw schema configuration for a connection |\n| `list_tables` | List all tables with enabled status and sync mode |\n| `get_table_columns` | Get column details for a specific table |\n| `reload_schema` | Refresh schema configuration from the source |\n\n## Limitations\n\nThe following capabilities are **not available** via Fivetran's REST API:\n\n| Capability | Status | Notes |\n|------------|--------|-------|\n| Column data types | Not available | Query your destination database directly for type info |\n| Sync history/logs | Not available | Fivetran uses [Log Service](https://fivetran.com/docs/logs) which streams to external destinations |\n| Detailed error logs for past syncs | Not available | Only current tasks/warnings are accessible |\n| Row-level sync status | Not available | Not exposed via API |\n| Data quality metrics | Not available | Not exposed via API |\n\nFor sync logs and historical data, consider setting up Fivetran's [Log Service](https://fivetran.com/docs/logs) to stream logs to your data warehouse.\n\n## Environment Variables\n\nThe server supports two naming conventions:\n\n| Preferred | Alternative |\n|-----------|-------------|\n| `FIVETRAN_API_KEY` | `FIVETRAN_SYNC_API_KEY` |\n| `FIVETRAN_API_SECRET` | `FIVETRAN_SYNC_API_SECRET` |\n\n## Development\n\n```bash\n# Clone the repository\ngit clone https://github.com/YimingYAN/fivetran-mcp.git\ncd fivetran-mcp\n\n# Install dependencies\nuv sync\n\n# Run the server locally\nuv run fivetran-mcp\n```\n\n## Cloud Deployment\n\n### Cloudflare Workers (Free)\n\nDeploy to Cloudflare Workers for a hosted MCP endpoint:\n\n```bash\ncd cloudflare\nuv tool install workers-py\nnpx wrangler login\nnpx wrangler secret put FIVETRAN_API_KEY\nnpx wrangler secret put FIVETRAN_API_SECRET\nuv run pywrangler deploy\n```\n\nSee [cloudflare/README.md](cloudflare/README.md) for detailed instructions.\n\n## License\n\n[MIT](LICENSE)\n",
  "bytes": 5778,
  "sha": "6b9c31d50302b41ab68d90437381571d85c63a288c19820415e65e6a3a38ebd7",
  "repo_slug": "yimingyan/fivetran-mcp",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_yimingyan_fivetran_mcp_ab9334f0/readme"
}