mcp-knowledgebase
Provides tools for querying and exploring MySQL databases, including schema discovery, table relationships, and safe read-only query execution with pagination.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@mcp-knowledgebase describe the users table"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP Knowledge Base Server
A lightweight Model Context Protocol (MCP) server for MySQL database querying and schema exploration. This tool enables AI agents to understand database structure and fetch data through a safe, read-only interface.
Features
Schema Discovery: Get complete database schema including tables, columns, data types, and comments
Key Relationships: Understand table relationships through primary keys, foreign keys, and unique keys
Safe Query Execution: Execute SELECT queries with automatic validation and pagination
Table Exploration: Search and describe tables easily
Related MCP server: MCP MySQL
Available Tools
Tool | Description |
| Get complete database schema for all tables |
| Get all keys and relationships for JOIN operations |
| Execute SELECT queries with pagination (read-only) |
| Get detailed information about a specific table |
| List all tables with metadata |
| Search tables and columns by pattern |
Quick Start
1. Install Dependencies
npm install2. Configure Environment
Copy .env.example to .env and update with your database credentials:
DB_HOST=127.0.0.1
DB_PORT=3306
DB_USER=admin
DB_PASSWORD=secret
DB_NAME=ez_ccsd1wy
DB_CONNECTION_LIMIT=10
QUERY_TIMEOUT_MS=30000
MAX_ROWS_PER_PAGE=1003. Run the Server
# Development mode (with auto-reload)
npm run dev
# Production mode
npm start4. Use with Claude/Cline
Add the server to your MCP configuration:
{
"mcpServers": {
"knowledgebase": {
"command": "node",
"args": ["/path/to/knowledgebase/src/index.js"],
"env": {
"DB_HOST": "127.0.0.1",
"DB_PORT": "3306",
"DB_USER": "admin",
"DB_PASSWORD": "secret",
"DB_NAME": "ez_ccsd1wy"
}
}
}
}Usage Examples
Get Database Schema
// Get all tables and their columns
await tools.get_schema()Get Table Relationships
// Understand how tables connect
await tools.get_keys()Execute a Query
// Simple SELECT query with pagination
await tools.raw_query({
query: "SELECT * FROM users WHERE active = 1",
page: 1,
limit: 50
})Describe a Table
// Get detailed table structure
await tools.describe_table({
tableName: "users"
})Search for Tables
// Find tables or columns
await tools.search_tables({
pattern: "user",
type: "all" // or "tables", "columns"
})Security
Read-Only Access: Only SELECT queries are allowed
Query Validation: All queries are validated before execution
Timeout Protection: Queries have a configurable timeout (default 30s)
Pagination: Automatic pagination prevents overwhelming responses
Project Structure
knowledgebase/
├── src/
│ ├── index.js # Main MCP server
│ ├── database.js # Database connection pool
│ └── tools/
│ ├── schema.js # Schema tools
│ └── query.js # Query tools
├── .env # Environment configuration
├── .env.example # Environment template
├── .nvmrc # Node.js version
├── package.json # Dependencies
└── README.md # This fileNode.js Version
This project uses Node.js 20.0.0. Use nvm to manage versions:
nvm useDocumentation
Agent Instructions - Guide for AI agents extending this tool
API Reference - Detailed tool specifications
License
ISC
Available Tools
6 toolsdescribe_tableA
Get detailed information about a specific table including all columns, their types, keys, and comments.
Inputs: tableName (required).
get_keysA
Get all keys (primary, foreign, unique) for all tables. This helps understand table relationships for JOIN operations. Returns primary keys, foreign keys, unique keys, and a relationships map showing how tables connect. Paginated to prevent token overflow.
Inputs: page, limit, keyType.
get_schemaA
Get the complete database schema for all tables. This includes table names, column names, data types, nullable status, defaults, and comments. Use this to understand the database structure before writing queries. Paginated to prevent token overflow.
Inputs: page, limit, compact.
get_tablesA
Get a list of all tables in the database with their metadata (type, comment, engine, row count). Paginated to prevent token overflow.
Inputs: page, limit, compact.
raw_queryA
Execute a raw SELECT query on the database. This tool provides read-only access with automatic pagination. The query is validated to ensure only SELECT statements are allowed. Results are paginated to prevent overwhelming responses.
Inputs: query (required), page, limit.
search_tablesA
Search for tables or columns by name pattern. Useful for finding specific tables or columns when you don't know the exact name. Paginated to prevent token overflow.
Inputs: pattern (required), page, type, limit, +1 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 6 tools
Tools mostly target distinct concerns: schema overview, keys, single-table details, table listing, search, and raw query. However, get_schema and describe_table overlap in providing column/key info, which could confuse an agent choosing between them. Overall boundaries are clear enough with careful reading.
Four tools use the get_ prefix (get_schema, get_keys, get_tables), but describe_table and search_tables use different verbs, and raw_query breaks the pattern entirely. The mix of get_, describe_, search_, and raw_ is readable but not consistent. A unified verb_noun convention would improve predictability.
Six tools is a well-scoped count for a database metadata and query server. Each tool covers a distinct aspect of exploration (schema, keys, tables, search, single-table details, raw queries), with no redundant tools. This is within the ideal range.
The toolset covers the main workflows: understanding schema, relationships, discovering tables, and running read-only queries. Minor gaps exist around indexes, views, or row sampling, but these are not critical for a knowledge base. The core surface is complete for navigation and querying.
Maintenance
Related MCP Connectors
PostgreSQL, MySQL, and SQL Server in one session. 26 read-only MCP tools for AI agents.
Safe, read-only Postgres and MySQL access for AI agents. Audit log + column-level controls.
Text-to-SQL MCP server: read-only queries on PostgreSQL, MySQL and SQL Server with your real schema
Read-only MCP tools for AI agent discovery, structured resources, and NIULAI information.
Related MCP Servers
- AlicenseAqualityDmaintenanceEnables AI agents to query and manage MySQL databases through a structured MCP interface, supporting SQL execution, table inspection, and database operations.911 npmMIT
- FlicenseAqualityBmaintenanceEnables AI assistants to connect to MySQL databases, execute read-only queries, list tables, and describe table schemas via MCP.101-
- AlicenseAqualityBmaintenanceProvides AI agents a secure MCP interface to a single MySQL or MariaDB database, enabling schema inspection, read-only queries, query planning, and guarded write operations with strict safety controls.6113 npmMIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to run read-only SQL against SQLite or MySQL databases through MCP tools for listing tables, describing schemas, and querying data. Wraps every call in guardrails such as statement allow-listing, row caps, query timeouts, PII masking, and audit logging.1MIT