================== /// MCP /// /// MCP /// ================== [server:online] [protocol:ready]
mcp-alchemy
by runekaagaard
MCP (Model Context Protocol) server that lets Claude Desktop explore, query and analyse SQL-Alchemy compatible databases (SQLite, PostgreSQL, MySQL/MariaDB, Oracle, MS-SQL, CrateDB, Vertica …).
279
50
Open Source01
all_table_names
Return all table names in the database
02
filter_table_names
Find and return table names that match a given substring
03
schema_definitions
Provide detailed schema information (columns, primary keys, foreign keys, nullability) for specified tables
04
execute_query
Execute a SQL query and return the results in a clean vertical format
Installation
1. Prerequisites
• Python ≥3.9 installed
• Access credentials for the databases you want the LLM to query (SQLite file path, or network host/port, user, password).
2. Clone or install
# Option A: install from PyPI (if/when published)
pip install mcp-alchemy
# Option B: install the latest code from GitHub
git clone https://github.com/runekaagaard/mcp-alchemy.git
cd mcp-alchemy
pip install -e .
3. Configure the server
Create a .env file or export environment variables:
MCP_HOST=0.0.0.0 # listening interface
MCP_PORT=8080 # listening port
DATABASE_URL="postgresql+psycopg2://user:password@host:5432/dbname" # SQLAlchemy URL for the primary database
# You can add multiple DATABASE_URL_n variables or a JSON/YAML config for several connections.
4. Launch the MCP server
python -m mcp_alchemy.server # or: mcp-alchemy serve
The server will start and expose an MCP endpoint (usually HTTP/JSON) that the LLM runtime can reach.
5. Verify
curl http://localhost:8080/health # should return OK
6. (Optional) Docker
docker build -t mcp-alchemy .
docker run -p 8080:8080 -e DATABASE_URL=... mcp-alchemy
API Specs
Documentation
License: Mozilla Public License 2.0
Updated 7/30/2025