MCP servers for databases
TL;DR
Database MCP servers give AI agents read or write access to a database through tools. Because live database credentials carry real risk, treat a database server as a read-only surface for agent callers by default and constrain its schema scope.
What to look for
Evaluate the connection mechanism, whether read-only mode exists, and what queries the exposed tools can run. Prefer a server whose tool surface you can bound to the minimum tables a task needs.
Read-only by default
Granting a model a write-capable database tool means any crafted query can change data. Prefer servers that support a read-only mode and limit the connected role to SELECT where possible.
In this collection
- Best MCP servers →
- How to choose an MCP server →
- Verified MCP servers →
- MCP server comparison methodology →
- MCP registries, directories, and marketplaces →
- Read-only database access with MCP →
- MCP servers for developers →
- MCP browser automation →
- MCP servers for search and research →
- MCP servers for productivity →
- MCP servers for DevOps and cloud →
- Enterprise MCP →
- Postgres MCP server guide →
- Playwright MCP →
- MCP server not showing in Claude Desktop →
- Cursor MCP tools not showing →
- MCP connection refused and timeout errors →
- Claude Desktop MCP configuration →
- VS Code MCP configuration →
- Remote MCP servers with OAuth →
- MCP server for GitHub →