Skip to main content
Glama
@arkty/redash-mcp

@arkty/redash-mcp

by arkty

@arkty/redash-mcp

Made with AI License: MIT npm

MCP Server for Redash — execute SQL queries, explore database schemas, and get results as JSON or CSV.

  • Ad-hoc queries without creating objects in Redash (no trace left)

  • SQL write guard — destructive operations (INSERT/UPDATE/DELETE/DROP) are blocked

  • CSV export — saves results to file, returns summary to LLM

  • Schema exploration — list tables, search columns, inspect structure

  • Default data source — configure once, query without specifying every time

Installation

npx -y @arkty/redash-mcp

Related MCP server: Generic Database MCP Server

Configuration

Claude Code (.mcp.json)

{
  "mcpServers": {
    "redash": {
      "command": "npx",
      "args": ["-y", "@arkty/redash-mcp"],
      "env": {
        "REDASH_URL": "https://your-redash.example.com",
        "REDASH_API_KEY": "your-api-key",
        "REDASH_DEFAULT_DATA_SOURCE_ID": "1"
      }
    }
  }
}

Environment Variables

Variable

Required

Default

Description

REDASH_URL

Yes

—

Redash instance URL

REDASH_API_KEY

Yes

—

User API key (from Redash profile settings)

REDASH_DEFAULT_DATA_SOURCE_ID

No

—

Default data source ID

REDASH_DEFAULT_TABLE

No

—

Default table name (shown in tool description)

REDASH_TIMEOUT

No

30000

Request timeout in ms

REDASH_MAX_ROWS

No

500

Max rows in JSON response

REDASH_POLL_INTERVAL

No

1000

Job polling interval in ms

REDASH_MAX_POLL_ATTEMPTS

No

120

Max poll attempts before timeout

REDASH_OUTPUT_DIR

No

.

Directory for CSV file output

REDASH_EXTRA_HEADERS

No

—

Extra HTTP headers (JSON or key=value;key2=value2)

Tools

execute_adhoc_query

Execute SQL directly against a data source. No query object is created in Redash. Destructive SQL is blocked.

Parameters: query, data_source_id?, format? ("json" | "csv")

list_data_sources

List all available data sources with IDs, names, and types.

get_data_source_schema

Get full database schema (all tables and columns) for a data source.

Parameters: data_source_id

get_table_info

Get column names and types for a specific table.

Parameters: data_source_id, table_name

search_schema

Search tables and columns by name pattern (case-insensitive).

Parameters: data_source_id, search

Security

  • Destructive SQL (INSERT, UPDATE, DELETE, DROP, ALTER, TRUNCATE, CREATE, etc.) is blocked at the MCP server level. The filter is intentionally simple — it checks for keyword presence in the query text. This means it may produce false positives (e.g., SELECT REPLACE(...) is blocked because REPLACE is a keyword). This is a safety net, not a bulletproof solution.

  • Use read-only database credentials in your Redash data source configuration — this is the real protection layer

  • API key is passed via environment variable, never hardcoded

License

MIT

Available Tools

5 tools
execute_adhoc_queryA
Read-only

Execute a SQL query directly against a data source. No query object is created in Redash — leaves no trace. Destructive SQL (INSERT/UPDATE/DELETE/DROP etc.) is blocked. For running queries this is the DEFAULT tool. Use create_query only when user explicitly asks to save a query to Redash web UI.

Inputs: query (required), format, output_path, data_source_id.

get_data_source_schemaA
Read-only

Get the full database schema (all tables and their columns) for a data source. Use list_data_sources first to find the data_source_id.

Inputs: data_source_id (required).

get_table_infoA
Read-only

Get detailed column information (names and types) for a specific table. Use search_schema if you don't know the exact table name.

Inputs: table_name (required), data_source_id (required).

list_data_sourcesA
Read-only

List all available Redash data sources (databases) with their IDs, names, and types

Inputs: no named parameters.

search_schemaA
Read-only

Search for tables and columns by name pattern (case-insensitive). Useful when you don't know the exact table or column name.

Inputs: search (required), data_source_id (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4.1/5.0

    Scored across 5 tools

    Disambiguation4/5

    The tools are mostly distinct: execute_adhoc_query handles query execution, list_data_sources handles data sources, and the remaining three handle schema discovery. There is some overlap among get_data_source_schema, get_table_info, and search_schema, but the descriptions clearly separate whole-schema, specific-table, and pattern-search use cases.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun snake_case pattern: execute_adhoc_query, list_data_sources, get_data_source_schema, get_table_info, and search_schema. The naming style is uniform and predictable.

    Tool Count5/5

    With 5 tools, the server is well-scoped for its apparent purpose: running ad-hoc SQL queries and exploring database schemas. Each tool has a clear role and no redundant tools are present.

    Completeness3/5

    The server covers ad-hoc query execution and schema discovery well, but it references create_query for saving queries to Redash without actually providing that tool. This leaves a notable gap for users who want to persist or manage queries in the Redash web UI.

    Maintenance

    ActivityNo data
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      B
      quality
      D
      maintenance
      Read-only MCP server wrapping the Redash REST API, enabling query listing, execution, data source inspection, and dashboard retrieval with optional per-user API key support.
      7
      -
    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables read-only SQL querying and schema inspection across MSSQL, PostgreSQL, and MySQL databases via MCP tools.
      ISC
    • F
      license
      Not graded
      quality
      D
      maintenance
      Enables running read-only SQL queries and exploring DuckDB databases through MCP tools like listing tables, describing schemas, and fetching paginated data.
      -
    • A
      license
      Not graded
      quality
      D
      maintenance
      Provides read-only access to Amazon Redshift warehouses through MCP, enabling listing databases, schemas, tables, and columns, running read-only queries, and explaining query plans.
      MIT