create_noteLABSCO SUMMARY
Table search, for schemas too big to list.
Listing every table is useless past a certain size, and that size arrives sooner than people expect. Pattern search and a table count turn discovery into a question rather than a wall of output — which is the difference between an assistant that can work in a large legacy schema and one that cannot.
What it is
Direct MySQL access with a discovery layer on top. Tables can be counted and searched by pattern rather than only listed, which is what makes it usable against a schema too large to read. Notes stored in the database are readable as MCP resources.
What you get
- Every table listed, and the total count
- Tables searched by a LIKE pattern, for finding one among hundreds
- The structure of a specific table
- Arbitrary SQL executed
- Notes created in the database, each with a title and content
- Those notes exposed as resources, addressable individually
Requirements
A running MySQL server and a database with the permissions you intend to use, its host, user, password and database name supplied through the environment. Node.js 18 or newer.
Setup effort
Build from source — git clone https://github.com/LeonMelamud/mysql-mcp.git
USE CASES
01
Finding the table you need in a large schema
Search by pattern instead of reading a list of hundreds, and check the count first to know what you are dealing with.
02
Reading what a table actually holds
Describe the table before writing anything against it, so the SQL matches the real column names and types.
03
Answering a data question in the conversation
Run the SQL and read the rows back where the question was asked, rather than moving to a client and back again.
04
Checking a query before it goes into code
Confirm the tables exist and the columns are named as you expect, then run the statement once to see what it returns.
05
Leaving a note where the next person will find it
Store findings as notes in the same database, addressable afterwards as resources — so what you worked out stays next to the data it is about.
6 tools
Listed in project docs. Your client may show a different list.
list_tablescount_tablessearch_tablesdescribe_tableexecute_sqlBEFORE YOU INSTALL
What to know first
Tools you get
6 tools named in the project's documentation.
Requires
- Node.js 18 or later
- MYSQL_PASSWORD in your environment
- MYSQL_USER in your environment
Where it runs
Both documented clients run the server on your machine.
INSTALLATION
Connect your client
You clone and build this one yourself.
Server setup
TERMINAL
git clone https://github.com/LeonMelamud/mysql-mcp.git
cd mysql-mcp
npm install
npm run buildMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"mysql": {
"command": "node",
"args": [
"<YOUR_PATH_TO_MYSQL_SERVER>/build/index.js"
],
"env": {
"MYSQL_HOST": "localhost",
"MYSQL_USER": "<YOUR_MYSQL_USER>",
"MYSQL_PASSWORD": "<YOUR_MYSQL_PASSWORD>",
"MYSQL_DATABASE": "<YOUR_MYSQL_DATABASE>"
}
}
}
}Cline
CLINE · JSON
{
"mcpServers": {
"mysql": {
"command": "node",
"args": [
"<YOUR_PATH_TO_MYSQL_SERVER>/build/index.js"
],
"env": {
"MYSQL_HOST": "localhost",
"MYSQL_USER": "<YOUR_MYSQL_USER>",
"MYSQL_PASSWORD": "<YOUR_MYSQL_PASSWORD>",
"MYSQL_DATABASE": "<YOUR_MYSQL_DATABASE>"
},
"disabled": false,
"autoApprove": []
}
}
}
