redash-mcp
This server is a read-only MCP wrapper for the Redash REST API (with the exception of saved query execution). Its capabilities include:
List & Search Queries (
list_queries): Browse or full-text search saved queries with pagination support.Get Query Details (
get_query): Retrieve detailed information about a specific query by ID (API keys are automatically redacted for security).Execute Saved Queries (
execute_query): Run a saved query by ID, optionally passing parameters and controlling cache behavior viamax_age; returns row data up to a configurable limit.List Data Sources (
list_data_sources): Retrieve all available Redash data sources.Inspect Data Source Schema (
get_data_source_schema): Fetch the schema (tables and columns) for a specific data source by ID.List & Search Dashboards (
list_dashboards): Browse or search dashboards with pagination support.Get Dashboard Details (
get_dashboard): Retrieve full details of a dashboard — including its widget configuration — using either its numeric ID or slug.
Provides read-only access to Redash REST API, including listing and searching queries, retrieving query details, executing saved queries with parameters, listing data sources and their schemas, and listing/retrieving dashboards.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@redash-mcp list my queries"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
redash-mcp
Redash REST API を読み取り専用でラップする MCP サーバ。唯一の例外として
保存済みクエリの実行(POST /api/queries/{id}/results)のみ許可する。
ツール
list_queries(q?, page, page_size)— クエリ一覧・検索get_query(query_id)— クエリ詳細(api_keyは秘匿)execute_query(query_id, parameters?, max_age)— クエリ実行 → 行データlist_data_sources()— データソース一覧get_data_source_schema(data_source_id)— スキーマlist_dashboards(q?, page, page_size)— ダッシュボード一覧・検索get_dashboard(identifier)— ダッシュボード詳細(id または slug)
Related MCP server: redash-mcp
設定(環境変数)
変数 | 必須 | 既定(プロセス) | 用途 |
| ✅ | — | 例 |
| stdioで必須 | — | 既定APIキー。HTTPは |
|
|
| |
|
| http バインドホスト。CLI | |
|
| http ポート。CLI | |
| (なし) | 許可 | |
|
| HTTP タイムアウト秒 | |
|
| ジョブ待機の上限秒 | |
|
| 返却行数の上限 |
実行
ローカル(uv):
uv run redash-mcp # stdio
uv run redash-mcp --transport streamable-http --host 127.0.0.1 --port 8000Docker(既定で streamable-http / 0.0.0.0:8000):
docker run --rm -p 8000:8000 \
-e REDASH_URL=http://redash -e REDASH_API_KEY=xxxx redash-mcpstreamable-http でネットワーク公開する場合は DNS リバインディング対策として REDASH_MCP_ALLOWED_HOSTS を設定することを推奨。
LibreChat 連携
各ユーザーが自分の Redash API キーを持てる(X-REDASH-API-KEY ヘッダ)。
REDASH_URL はサーバ側で固定し、REDASH_API_KEY(サーバ既定キー)は設定しない。
各リクエストの X-REDASH-API-KEY ヘッダを実効キーとする。
1. LibreChat の mcpServers 設定
redash:
type: "streamable-http"
url: "http://redash-mcp:8000/mcp"
headers:
X-REDASH-API-KEY: "{{REDASH_API_KEY}}"
customUserVars:
REDASH_API_KEY:
title: "Redash API Key"
description: "Redash の [ユーザー設定] から発行した API キー"2. docker-compose(LibreChat と同一ネットワークに接続)
コンテナは LibreChat と同じ Docker ネットワークに置き、REDASH_MCP_ALLOWED_HOSTS
で許可 Host を絞る。ホストへのポート公開は不要(LibreChat から内部ネットワーク
経由で http://redash-mcp:8000/mcp に接続する)。
services:
redash-mcp:
image: redash-mcp
build: .
environment:
REDASH_URL: http://redash
REDASH_MCP_ALLOWED_HOSTS: "redash-mcp"
# REDASH_API_KEY は設定しない(各ユーザーが X-REDASH-API-KEY ヘッダで供給)
networks:
- librechat_default
networks:
librechat_default:
external: trueスタンドアロン(LibreChat を使わない)構成は
docker-compose.example.ymlを参照。
開発
uv sync
uv run pytestAvailable Tools
7 toolsexecute_queryB
保存済みクエリを実行し行データを返す。max_age=0でキャッシュ回避。
Inputs: query_id (required), max_age, parameters.
get_dashboardC
ダッシュボード1件(id または slug)。ウィジェット構成を含む。
Inputs: identifier (required).
get_data_source_schemaC
データソースのスキーマ(テーブル・カラム)。
Inputs: data_source_id (required).
get_queryC
指定IDのクエリ詳細(api_keyは秘匿済み)。
Inputs: query_id (required).
list_dashboardsC
ダッシュボード一覧・検索。
Inputs: q, page, page_size.
list_data_sourcesC
データソース一覧。
Inputs: no named parameters.
list_queriesB
クエリ一覧・検索。qで全文検索、page/page_sizeでページング。
Inputs: q, page, page_size.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 7 tools
Each tool targets a distinct resource and action: queries (execute vs. get details vs. list), dashboards (get vs. list), and data sources (list vs. get schema). No overlapping purposes.
All tool names follow a consistent verb_noun pattern (e.g., execute_query, get_dashboard, list_queries), with uniform use of snake_case.
Seven tools is appropriate for a Redash MCP server, covering essential read and list operations without unnecessary bloat.
The tool set lacks any create, update, or delete operations for queries, dashboards, or data sources, leaving significant gaps in lifecycle management. An agent cannot create or modify resources.
Related MCP Connectors
Read-only MCP server for AIStatusDashboard status, incidents, metrics, and fallback recommendations.
Read-only MCP server for The Quiet Protocol's engines, benchmarks, proof, and business data.
OAuth-protected, read-only-by-default MCP server for provenance-labeled QuillCaddie project memory.
Read-only MCP server for ClassQuill, a tutoring-business-management platform.
Related MCP Servers
- AlicenseBqualityBmaintenanceModel Context Protocol (MCP) server that integrates Redash with AI assistants like Claude, allowing them to query data, manage visualizations, and interact with dashboards through natural language.674,448 npm108MIT
- AlicenseAqualityDmaintenanceModel Context Protocol (MCP) server for Redash - manage queries, dashboards, and visualizations through AI assistants like Claude.699 PyPIMIT
- AlicenseNot gradedqualityFmaintenanceRead-only MCP server for SQL databases (SQL Server, Postgres, SQLite) with multi-server support and three-layer safety using AST validation and linting.MIT
- AlicenseNot gradedqualityDmaintenanceMCP server that connects AI assistants to Redash, enabling listing of assets, read-only query execution, dashboard inspection, and alert management through the Redash API.MIT