listLABSCO SUMMARY
Legacy Access files, opened up to a conversation.
Access databases are usually where data goes to become unreachable. This reaches them with ordinary SQL, moves rows in and out through CSV and Excel, and lets you leave notes on a file so the next person does not have to guess at the schema.
What it is
An MCP server for Microsoft Access (.mdb, .accdb) and SQLite 3 (.db, .sqlite, .sqlite3) files. It opens a database file - or an in-memory one - runs SQL against it for reads and writes, imports and exports CSV, imports from Excel, and keeps human-readable notes alongside a file so the next session knows what is in it.
What you get
- CSV imported into a database table
- A table exported to CSV
- An Excel sheet imported into a table
- Notes written about a database file, or linked to it
- Those notes read back, or discovered across a directory
- SQL run against Access and SQLite alike, for both retrieval and updates
- Databases opened read-only when you want them safe, or in memory when you want them temporary
Requirements
uv, with fastmcp, pandas, sqlalchemy-access and openpyxl pulled in on the command line; the repository is cloned rather than installed from a package index. The server can read and modify any data in a database it connects to, so open with readOnly=true unless you mean to write - note that read-only is not available for in-memory SQLite. Excel import works; Excel export is not implemented.
Setup effort
Build from source
USE CASES
01
Getting data out of an old Access file
Export the table straight to CSV instead of installing Access to look at it.
02
Loading a spreadsheet someone emailed into a real table
Import the Excel sheet directly, without a CSV conversion step in between.
03
Asking questions of a legacy database without Access installed
Connect to the .mdb file with readOnly=true and run SELECT statements against it, so a mistake cannot change anything.
04
Leaving a note about what a file actually contains
Write the note against the database, and read it back at the start of the next session.
05
Making sense of a folder of database files you inherited
Notes can be discovered across a directory, so one call summarises the whole folder.
06
Moving a table between two databases
Export from one and import into the other; the CSV is the bridge.
11 tools
Listed in project docs. Your client may show a different list.
createconnectdisconnectqueryupdateimport_csvexport_csvimport_excelread_noteswrite_notesBEFORE YOU INSTALL
What to know first
Tools you get
11 tools named in the project's documentation.
Requires
uv on PATH
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
Server setup
TERMINAL
git clone https://github.com/scanzy/mcp-server-access-mdb.gitMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"access-mdb": {
"command": "uv",
"args": [
"run",
"--with",
"fastmcp",
"--with",
"pandas",
"--with",
"sqlalchemy-access",
"--with",
"openpyxl",
"fastmcp",
"run",
"<YOUR_PATH_TO_REPO>/server.py"
]
}
}
}<YOUR_PATH_TO_REPO>/server.py is the full path to server.py in the folder the clone created (run pwd inside it and add /server.py). The client does not start the server from inside your clone, so the relative path as printed will not be found.
