MCP servers that inspect database schemas
List tables, describe columns, and explore schema so the AI writes correct queries.
9 servers · Last updated June 17, 2026
TL;DR: Before an agent can query well, it needs to see the schema. These servers list tables, describe columns and types, and surface relationships — the context that turns vague prompts into correct SQL. Pair one with an execute-SQL server for an end-to-end data agent.
Bottom line: if you only try one, Postgres MCP Pro is the most popular, verified option for this (2,900★). 8 more compared below.
Compare 9 servers
| Server | Transport | Auth | Verified | Stars | Tools for this |
|---|---|---|---|---|---|
| Postgres MCP Pro | Local (stdio) | API key | 2,900 | list_schemas, list_objects | |
| Supabase MCP Server | Remote (HTTP) | OAuth | 2,700 | list_tables | |
| Neon MCP Server | Remote (HTTP) | OAuth | 2,100 | list_projects / describe_project / create_project / delete_project, create_branch / delete_branch / describe_branch / reset_from_parent, get_database_tables / describe_table_schema +1 | |
| MongoDB MCP Server | Local (stdio) | API key | 1,500 | collection-schema | |
| Neo4j Cypher MCP Server | Local (stdio) | API key | 1,300 | get_neo4j_schema | |
| ClickHouse MCP Server | Local (stdio) | API key | 1,000 | list_databases, list_tables | |
| MotherDuck / DuckDB MCP Server | Local (stdio) | API key | 800 | list_databases, list_tables | |
| Chroma MCP Server | Local (stdio) | API key | 600 | chroma_list_collections | |
| Pinecone Developer MCP Server | Local (stdio) | API key | 500 | describe-index, describe-index-stats |
The servers
Read/write Postgres access plus index tuning, EXPLAIN plans, and database health analysis for AI agents.
Official Supabase server: manage tables, run SQL, branches, configs and edge functions from your AI client.
Manage serverless Postgres on Neon with natural language: projects, branches, migrations, and SQL.
Official MongoDB server connecting Atlas, Community, or Enterprise deployments to AI tools.
Run Cypher queries and explore graph schema on Neo4j from your AI assistant.
Official ClickHouse server: read-only SQL queries plus database/table exploration and chDB support.
Query local DuckDB files, in-memory DBs, S3, or MotherDuck cloud with read/write SQL.
Official Chroma server: create collections and run vector, full-text, and metadata search.
Official Pinecone server: manage indexes, upsert/search records, rerank, and search Pinecone docs.
Use these in a stack
FAQ
Why does schema inspection matter for AI agents?
LLMs hallucinate table and column names. Giving the agent a tool to list tables and describe schema first dramatically cuts wrong-query errors.
Do these work with NoSQL too?
Some do — document stores expose collection/index listing tools that play the same role as table introspection in SQL databases.