redash-mcp

Model Context Protocol (MCP) server for Redash - manage queries, dashboards, and visualizations through AI assistants like Claude.

Features

• 7 tools, 30 actions - compressed for minimal context usage
• Full query management (list, search, create, update, archive, delete, run, adhoc, export, schedule)
• Dashboard management (list, get, create, publish, delete)
• Widget management with positioning (add, move, delete)
• Alert management (list, get, create, update, delete)
• Visualization creation (pie, line, bar, counter charts)
• Data source listing

Installation

pip install redash-mcp

Or with uvx:

uvx redash-mcp

Configuration

Environment Variables

Claude Code

Add to ~/.claude.json (user-level config):

{
  "mcpServers": {
    "redash": {
      "type": "stdio",
      "command": "uvx",
      "args": ["redash-mcp"],
      "env": {
        "REDASH_URL": "https://your-redash-instance.com",
        "REDASH_API_KEY": "your-api-key"
      }
    }
  }
}

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "redash": {
      "command": "uvx",
      "args": ["redash-mcp"],
      "env": {
        "REDASH_URL": "https://your-redash-instance.com",
        "REDASH_API_KEY": "your-api-key"
      }
    }
  }
}

Or if installed via pip:

{
  "mcpServers": {
    "redash": {
      "command": "redash-mcp",
      "env": {
        "REDASH_URL": "https://your-redash-instance.com",
        "REDASH_API_KEY": "your-api-key"
      }
    }
  }
}

Tools

redash_data_sources

List all available data sources.

redash_query

Manage Redash queries.

redash_dashboard

Manage Redash dashboards.

redash_widget

Manage dashboard widgets.

redash_alert

Manage query alerts.

redash_viz

Create visualizations.

Note: For multiple Y columns, pass comma-separated values: y="count,total,avg"

Examples

Create a dashboard with visualizations

1. redash_data_sources() → get data_source_id
2. redash_query(action="create", name="Daily Stats", query="SELECT ...", data_source_id=1)
3. redash_viz(type="line", query_id=123, name="Trend", x="date", y="count")
4. redash_dashboard(action="create", name="My Dashboard")
5. redash_widget(action="add", dashboard_id=456, viz_id=789)
6. redash_dashboard(action="publish", id=456)

Run ad-hoc query

redash_query(action="adhoc", query="SELECT COUNT(*) FROM users", data_source_id=1)

Export query results

redash_query(action="export", id=123, path="/tmp/results.csv")
redash_query(action="export", id=123, path="/tmp/results.json")

Search and update query

redash_query(action="search", q="daily")
redash_query(action="update", id=123, query="SELECT ... WHERE date > NOW() - INTERVAL '7 days'")

Python Library Usage

You can also use redash-mcp as a Python library:

import os
os.environ["REDASH_URL"] = "https://your-redash.com"
os.environ["REDASH_API_KEY"] = "your-key"

from redash_mcp import (
    list_queries, create_query, run_query,
    create_dashboard, publish_dashboard,
    line, bar, pie, counter,
    add_widget
)

# Create query
q = create_query("My Query", "SELECT * FROM events", data_source_id=1)

# Create visualization
viz = line(q["id"], "Events Trend", x="date", y=["count"])

# Create dashboard and add widget
d = create_dashboard("My Dashboard")
add_widget(d["id"], viz["id"])
publish_dashboard(d["id"])

Why redash-mcp?

• Context efficient - Only 7 tools (~500 tokens) with 30 actions
• Full-featured - Queries, dashboards, widgets, and visualizations
• Production ready - Proper error handling and timeouts
• Dual use - Works as MCP server and Python library

License

MIT