Performs SQL CRUD operations using AI prompts and FastMCP
Steps to run the MCP server :-
Prerequisites :-
Download Gemini Code Assist VS Code extension
Make sure that the gemini code extension is in agent mode
IMPORTANT - Reload VS Code after updating any settings to refresh
Use /mcp command to list all tools in chat
Add this to .gemini/settings.json in C:/Users/Lenovo/.gemini:
{
"mcpServers": {
"my-persistent-memory": {
"command": "C:/Users/Lenovo/Downloads/mcp_sql_project/venv/Scripts/python.exe",
"args": ["C:/Users/Lenovo/Downloads/mcp_sql_project/server.py"],
"env": {
"PYTHONPATH": "C:/Users/Lenovo/Downloads/mcp_sql_project"
}
}
}
}
Prompts:-
Test the full "CRUD" (Create, Read, Update, Delete) cycle:
Store: "Add key as project lead and content is Alex to db." (Uses store_memory)
Retrieve: "Who is my project lead?" (Uses retrieve_memory)
Fetch all: "List all memories" (Uses list_memories)
Update: "Actually, Alex left; update my project lead memory to say the lead is now Sarah." (Uses update_memory)
Delete: "I'm done with the project lead info, you can delete that memory now." (Uses delete_memory)