home / mcp / databases mcp server
Provides an MCP server to interact with Microsoft Access and SQLite 3 databases, with CSV/Excel import/export and file notes.
Configuration
View docs{
"mcpServers": {
"carltk-mcp-server-access-mdb-with-access97": {
"command": "uv",
"args": [
"run",
"--with",
"fastmcp",
"--with",
"pandas",
"--with",
"sqlalchemy-access",
"--with",
"openpyxl",
"fastmcp",
"run",
"path/to/repo/server.py"
]
}
}
}You can use this MCP server to let AI interact with Microsoft Access and SQLite 3 databases, with support for importing and exporting CSV and Excel data and storing human-readable notes about files. It emphasizes easy experimentation with database files while warning you to handle data carefully due to full access rights.
You work with an MCP client to connect to the Databases MCP Server. This server provides tools to manage databases, run queries, and import or export data from common formats. Start the server locally and point your MCP client to the available connection method. Use the client to connect to an Access or SQLite database, run SQL queries, and manage notes associated with files.
A typical workflow includes connecting to a database file, inspecting tables, executing read or write queries, and importing data from CSV or Excel files. Remember that Access 97 databases are read-only, while modern Access and SQLite databases support full read/write operations through the server.
Prerequisites: you need a runtime environment that can execute MCP servers (such as uv), and you will configure the server through a JSON configuration snippet. Ensure you have access to the database files you want to work with.
1) Prepare the configuration with your MCP host. 2) Run the MCP server using the exact command and arguments shown in the configuration. 3) Connect your MCP client to the server and begin using the available tools.
Note: The following configuration enables the server to run with necessary components like fastmcp, pandas, and sql tooling for Access databases, and opens the path to the server runner.
Supported database types include Microsoft Access (.mdb and .accdb) with modern formats supported via ACE ODBC, legacy Access 97 via a read-only parser, SQLite 3 (.db, .sqlite, .sqlite3), and in-memory SQLite when no database path is provided.
Access 97 support is read-only and limited to basic SELECTs without complex joins or subqueries. Parameterized queries are not supported for Access 97. The server automatically detects Access 97 and uses the read-only path when needed.
Available tools include managing databases and data, as well as handling notes. For example, you can list databases, connect to or create a database, run queries, import from CSV, export to CSV, and manage notes about files.
This server has access to read and modify database content. Use it with care to avoid unintended data changes or losses.
{
"mcpServers": {
"access_mdb": {
"command": "uv",
"args": [
"run",
"--with", "fastmcp",
"--with", "pandas",
"--with", "sqlalchemy-access",
"--with", "openpyxl",
"fastmcp", "run",
"path/to/repo/server.py"
],
}
}
}List all active databases available in the server.
Create a new database file (for Microsoft Access, copies the empty.mdb template).
Connect to an existing database file, or create an in-memory database if the file is not specified.
Close a database connection. For in-memory databases, this clears all data.
Execute a SQL query to retrieve data from a database.
Execute a SQL query to insert/update/delete data in a database.
Imports data from a CSV file into a database table.
Exports data from a database table to a CSV file.
Imports data from an Excel file into a database table.
Reads notes from a specified file or discovers notes within a directory.
Writes notes to a specified file or links notes to a database.