Official
Query and inspect MongoDB collections from Claude, Cursor, and VS Code.
database, nosql, mongodb, atlas
Best Databases MCP Servers in 2026
Top database MCPs for Claude and Cursor: Postgres, SQLite, Supabase, and more. Inspect schemas and run queries safely from AI agents — verified for 2026.
About Databases MCP servers
The best MCP servers for databases in 2026 are Postgres MCP Pro for self-hosted Postgres, the Supabase MCP for Supabase-hosted Postgres, and SQLite for local prototyping — then add the warehouse or store that matches your stack (BigQuery, Snowflake, ClickHouse, MongoDB, MySQL, Redis, DuckDB, or Elasticsearch). Database MCP servers ship with read-only modes, server-side row limits, and explicit table allowlists so the agent can introspect the schema, draft SQL, and return rows — without putting production at risk.
Choose by stack and risk profile. If you use Supabase, the official Supabase MCP wins because it understands row-level security and project scoping. For self-hosted Postgres, Postgres MCP Pro is the right default — the reference modelcontextprotocol/server-postgres was archived on 2025-07-10 with a Datadog-documented read-only bypass and will not receive further fixes. SQLite is the fastest path for local prototyping — no server process, no network. For warehouse work, BigQuery, Snowflake, and ClickHouse each have a maintained MCP; DuckDB covers the local CSV/Parquet case; MongoDB, MySQL, and Redis MCPs cover the document store, OLTP, and cache lanes. Elasticsearch sits in its own seat for log and search debugging.
Common mistakes to avoid: connecting an MCP with read-write credentials when read-only would do, pointing it at production from a development client, and skipping query-budget limits — agents will happily emit a SELECT * on a billion-row table. Each MCP below documents auth model (api-key, OAuth, password) and recommended scoping. Start with read-only against a non-production database, run a few schema-introspection prompts, then graduate to read-write only after you trust the workflow.
All Databases MCPs
11 MCPs ranked by popularity. Filter by attribute or search by name.
11 of 11 MCPs
| # | MCP | Tags | Setup | Complexity | Labels | |
|---|---|---|---|---|---|---|
| 1 | MongoDB Query and inspect MongoDB collections from Claude, Cursor, and VS Code. | database, nosql | 5 min | Low | Official | |
| 2 | Redis Inspect keys, run commands, and debug Redis from an AI agent. | database, cache | 3 min | Low | Official | |
| 3 | SQLite Local SQLite database access with full read/write support. | database, sqlite | 2 min | Low | ||
| 4 | ClickHouse Query ClickHouse and ClickHouse Cloud columnar warehouses from an AI agent. | database, warehouse | 10 min | Medium | Official | |
| 5 | Elasticsearch Search indices, inspect mappings, and run queries against Elasticsearch from an AI agent. | database, search | 10 min | Medium | Official | |
| 6 | DuckDB In-process analytical SQL over files and Parquet from an AI agent. | database, olap | 2 min | Low | ||
| 7 | Postgres MCP Pro Production-grade Postgres MCP with index advice, EXPLAIN, and health checks. | database, sql | 5 min | Medium | ||
| 8 | MySQL Run read-only MySQL queries and inspect schemas from an AI agent. | database, sql | 5 min | Low | ||
| 9 | BigQuery Query Google BigQuery datasets from Claude, Cursor, and VS Code. | database, warehouse | 10 min | Medium | ||
| 10 | Snowflake Inspect schemas and run SQL against a Snowflake account from an AI agent. | database, warehouse | 15 min | Medium | ||
| 11 | Supabase Full Supabase access: database, auth, storage, and edge functions. | supabase, database | 5 min | Low |
Choose the right MCP
Quick decision guide based on your use case.
| If you need… | Start with |
|---|---|
| You use Supabase | Use the Supabase MCP |
| You run self-hosted Postgres | Use Postgres MCP Pro |
| You need local, serverless storage | Use the SQLite MCP |
| Your stack is MongoDB | Use the MongoDB MCP |
| Your stack is MySQL or MariaDB | Use the MySQL MCP |
| You need to debug Redis or Valkey | Use the Redis MCP |
| You query local files (CSV, Parquet) | Use the DuckDB MCP |
| Your warehouse is BigQuery | Use the BigQuery MCP |
| Your warehouse is Snowflake | Use the Snowflake MCP |
| Your warehouse is ClickHouse | Use the ClickHouse MCP |
| You search logs with Elasticsearch | Use the Elasticsearch MCP |
Top Databases MCPs ranked
Detailed cards with setup time, complexity, and key labels.
Official
Inspect keys, run commands, and debug Redis from an AI agent.
database, cache, redis, valkey
Official
Query ClickHouse and ClickHouse Cloud columnar warehouses from an AI agent.
database, warehouse, clickhouse, olap
Official
Search indices, inspect mappings, and run queries against Elasticsearch from an AI agent.
database, search, elasticsearch, elastic
In-process analytical SQL over files and Parquet from an AI agent.
database, olap, duckdb, parquet
Production-grade Postgres MCP with index advice, EXPLAIN, and health checks.
database, sql, postgres, performance
Run read-only MySQL queries and inspect schemas from an AI agent.
database, sql, mysql, mariadb
Query Google BigQuery datasets from Claude, Cursor, and VS Code.
database, warehouse, bigquery, google-cloud
Inspect schemas and run SQL against a Snowflake account from an AI agent.
database, warehouse, snowflake, sql
11
Full Supabase access: database, auth, storage, and edge functions.
supabase, database, backend, postgres
Archived (historical reference)
1 Databases entry is archived — the upstream package was deprecated or pulled, or a documented security issue applies. The detail page is preserved for historical reference and migration guidance, but these are NOT current editorial picks.
Also in the ChatGPT Apps directory
1 Databases brand is available as one-click ChatGPT Apps. These are not ranked alongside the editorial picks above — they're listings from chatgpt.com/apps surfaced here for brand-search continuity.
FAQ: Databases MCPs
Which database MCP should I start with?
Match it to your stack: Supabase → Supabase MCP, self-hosted Postgres → Postgres MCP Pro, local prototyping → SQLite, BigQuery → BigQuery MCP, Snowflake → Snowflake MCP, ClickHouse → ClickHouse MCP, MongoDB → MongoDB MCP, MySQL or MariaDB → MySQL MCP, Redis cache debugging → Redis MCP, local CSV/Parquet analytics → DuckDB, log search → Elasticsearch MCP.
Are database MCPs safe for production?
Only with read-only credentials and the MCP-level read-only flag both enabled. Treat the MCP read-only flag as advisory: also create a SELECT-only database user (or a viewer-role API key for warehouses) so the MCP cannot mutate state even if a tool surface ever changes.
Why is the reference PostgreSQL MCP archived?
Anthropic archived github.com/modelcontextprotocol/servers on 2025-07-10. Datadog Security Labs has documented a SQL-injection issue that escapes the server-side read-only restriction. Migrate to Postgres MCP Pro (self-hosted) or the Supabase MCP (Supabase-hosted) and grant the connecting Postgres role only SELECT as defence in depth.
How do I avoid burning warehouse credits?
For BigQuery, always run dry_run_query before query — it returns bytes-processed before you spend. For Snowflake, scope the MCP to a dedicated XS warehouse with a resource monitor. For ClickHouse Cloud, set max_execution_time on the user profile.