redash-mcp

English | 한국어 | 日本語

MCP server that connects Redash to Claude AI — query data, manage dashboards, and run SQL with natural language.

Why · Features · Install · Env · Examples · Privacy

Why redash-mcp?

There are several Redash MCP servers. This one is built for letting an LLM touch production data safely:

• 🛡️ SQL Safety Guard — blocks DROP/TRUNCATE/ALTER and DELETE/UPDATE without WHERE; strict/warn/off modes, PII detection, and auto-LIMIT. Hand Claude your real Redash without flinching.
• 🧠 BIRD Smart Query — analyzes the question, auto-selects the right tables, and guides SQL generation (based on the BIRD text-to-SQL methodology). Optional Claude Haiku fallback for table selection.
• ⚡ One-command setup — npx redash-mcp setup configures Claude Desktop / Claude Code for you. No hand-editing JSON.
• 🔒 Fully local — talks directly to your Redash instance. API key and query results never leave your machine.
• 📊 End-to-end — query, save, fork, dashboards, widgets, and alerts — 20+ tools across 6 categories.

Features

Tools

SQL Safety Guard

Protects your database from dangerous queries:

• Blocked always: DROP, TRUNCATE, ALTER TABLE, GRANT/REVOKE, DELETE/UPDATE without WHERE
• Warned (warn mode) / Blocked (strict mode): SELECT *, queries without WHERE or LIMIT, PII column access
• Auto-LIMIT: Automatically appends LIMIT N when REDASH_AUTO_LIMIT is set

Query Cache

Results are cached in-memory to reduce redundant API calls:

• TTL: configurable via REDASH_MCP_CACHE_TTL (default: 300s)
• Max memory: configurable via REDASH_MCP_CACHE_MAX_MB (default: 50MB)

Installation

Auto Setup (Recommended)

npx redash-mcp setup

The setup wizard will guide you through configuring Claude Desktop, Claude Code (CLI), or both.

Shell Script Install

Installs Node.js, Claude Desktop, and MCP config all at once:

curl -fsSL https://raw.githubusercontent.com/seob717/redash-mcp/main/install.sh | bash

Manual Setup

1. Get your Redash API Key

Go to Redash → Profile (top right) → Edit Profile → Copy API Key

2-A. Claude Desktop

Open ~/Library/Application Support/Claude/claude_desktop_config.json and add:

{
  "mcpServers": {
    "redash-mcp": {
      "command": "npx",
      "args": ["-y", "redash-mcp"],
      "env": {
        "REDASH_URL": "https://your-redash-instance.com",
        "REDASH_API_KEY": "your_api_key_here"
      }
    }
  }
}

Fully quit and restart Claude Desktop after saving.

2-B. Claude Code (CLI)

Open ~/.claude/settings.json and add:

{
  "mcpServers": {
    "redash-mcp": {
      "command": "npx",
      "args": ["-y", "redash-mcp"],
      "env": {
        "REDASH_URL": "https://your-redash-instance.com",
        "REDASH_API_KEY": "your_api_key_here"
      }
    }
  }
}

macOS: If npx is not found, run which npx to get the full path and use that instead.

Environment Variables

Required

Optional

Usage Examples

Just ask Claude in natural language:

• "Show me the columns in the users table"
• "Run a query to get order counts for the last 7 days"
• "List all saved queries"
• "Show widgets in the revenue dashboard"
• "Create an alert when daily signups drop below 100"

Example 1: Query data with natural language

Prompt: "How many new users signed up this month?"

Tool flow:

1. list_data_sources → Identify the target data source
2. smart_query → Analyze the question, auto-select the User table, provide SQL generation guidance
3. run_query → Execute the generated SQL

Result:

There were 18,197 new signups this month.

Example 2: Complex business questions

Prompt: "What percentage of last week's new users made a purchase?"

Tool flow:

1. smart_query → Analyze the question, auto-select User and Payment tables, provide JOIN query guidance
2. run_query → Execute the SQL

Result:

Out of 1,204 new users last week, 312 made a purchase (25.9%).

Example 3: Create a query and dashboard

Prompt: "Create a monthly revenue trend query and add it to a dashboard"

Tool flow:

1. smart_query → Analyze revenue-related tables
2. create_query → Save the "Monthly Revenue Trend" query
3. create_dashboard → Create a "Revenue Dashboard"
4. get_query → Get the visualization ID from the saved query
5. add_widget → Add the chart widget to the dashboard

Result:

Created "Revenue Dashboard" with the monthly revenue trend chart.
View in Redash: https://your-redash.com/dashboard/monthly-revenue

Privacy Policy

Data Collection and Processing

redash-mcp is a local MCP server that communicates directly with your Redash instance. No intermediate servers are involved.

Third-Party Sharing

We do not sell or share user data with third parties. When the LLM Fallback feature is active, only table name lists are sent to the Anthropic API, and only when the user has explicitly configured an ANTHROPIC_API_KEY.

Data Retention

• Config files: Stored locally in ~/.redash-mcp/ (user can delete at any time)
• Query cache: In-memory only, cleared on server shutdown
• Schema cache: In-memory only, auto-expires after 10-minute TTL

Contact

For inquiries and security reports: GitHub Issues

Disclaimer

redash-mcp is an unofficial, community-built integration and is not affiliated with, endorsed by, or sponsored by Redash. "Redash" is a trademark of its respective owners and is used here only to describe compatibility. This project communicates with Redash solely through its public REST API.

License

MIT © seob717