Full comparison table

All scores are on a 0–100 scale. The average across 720 data tools is 27.4. The tools below represent the top tier — all scoring above 70.

Top 10 breakdown

#1 — motherduckdb/mcp-server-motherduck (89.06)

motherduckdb/mcp-server-motherduck leads at 89.06. DuckDB runs in-process — there's no separate server, no network round-trips for local queries. This makes it the fastest analytical option for agents processing local Parquet files, CSV exports, or in-memory datasets. When local scale isn't enough, MotherDuck extends it to cloud-hosted DuckDB. 80% issue close rate. Best-in-category for OLAP and analytical workloads.

#2 — redis/mcp-redis (88.84)

redis/mcp-redis scores 88.84 as the official Redis Labs server. For data pipelines, Redis is the glue — pipeline state, task queues, pub/sub messaging, rate limiting, and stream processing. This server exposes all of it through natural language: strings, hashes, sets, sorted sets, streams, and full-text search. 68% issue close rate, official vendor backing for long-term reliability.

#3 — mongodb-js/mongodb-mcp-server (88.56)

mongodb-js/mongodb-mcp-server is the official MongoDB server at 88.56. With 965 stars and an 84% issue close rate — the best of the top 3 — it's the most responsive to issues. Covers both local instances and Atlas cloud clusters. For pipelines ingesting semi-structured data (logs, events, API responses), MongoDB remains the dominant document store and this is its reference MCP server.

#4 — tableau/tableau-mcp (84.32)

tableau/tableau-mcp is Tableau's official server at 84.32. It's the bridge between AI agents and business intelligence — agents can query workbooks, extract underlying data, navigate dashboards, and understand what your organization is measuring. 202 stars, 73% issue close rate. Essential for teams where Tableau sits at the end of the data pipeline as the consumption layer.

#5 — benborla/mcp-server-mysql (85.22)

benborla/mcp-server-mysql scores 85.22 with 1,361 stars — the most-starred MySQL MCP server in the index. It's read-only by design, which is a deliberate safety constraint for production pipelines. Data engineers use it for schema inspection, data quality checks, and SELECT-based validation without risk of accidental mutation. If you need write access, build on top of it or pair with a write-capable layer.

#6 — astronomer/agents (77.89)

astronomer/agents scores 77.89 as the official Astronomer (Apache Airflow) toolkit for data engineering agents. It enables AI-assisted DAG authoring, pipeline debugging, and workflow orchestration — the full Airflow development loop. 280 stars, 57% issue close rate. The highest-scoring Airflow-adjacent tool in the index and the natural choice for teams running Airflow on Astronomer's managed platform.

#7 — dbt-labs/dbt-mcp (76.04)

dbt-labs/dbt-mcp is the official dbt Labs server at 76.04. It sits between your dbt project and any agent — run models, execute tests, inspect schemas, and query outputs through natural language. 507 stars, 68% issue close rate. It's the only dbt-native MCP tool in the index. If your transformation layer is dbt, this is required.

#8 — runekaagaard/mcp-alchemy (74.93)

runekaagaard/mcp-alchemy scores 74.93 with a 100% issue close rate — every reported issue resolved. It uses SQLAlchemy under the hood, which means it connects to any SQLAlchemy-compatible database: Postgres, MySQL, SQLite, MSSQL, Oracle, and more. 397 stars. The best single tool for polyglot data environments where you're querying multiple engines from one pipeline.

#9 — LucasHild/mcp-server-bigquery (75.28)

LucasHild/mcp-server-bigquery scores 75.28 and is the primary MCP tool for Google BigQuery. It supports SQL query execution, dataset inspection, and table schema exploration. 123 stars, 41% issue close rate — the lowest in this list, which signals room for improvement. But it's the only mature BigQuery MCP option in the index, making it the default for GCP-based data warehouses.

#10 — yangkyeongmo/mcp-server-apache-airflow (71.10)

yangkyeongmo/mcp-server-apache-airflow rounds out the list at 71.10. It wraps the Airflow REST API — trigger DAGs, check run status, monitor task logs, manage pipeline health. 148 stars, 56% issue close rate. The choice for teams running self-managed Apache Airflow (versus Astronomer's managed platform). Complements astronomer/agents for Airflow-native orchestration use cases.

Choosing by stack

Scoring methodology

AgentRank scores combine five GitHub signals: stars (15%), commit freshness (25%), issue close rate (25%), contributor count (10%), and inbound dependents (25%). Scores are updated nightly from live crawl data. The average score across all 720 data-category tools is 27.4. Full methodology →