A MCP server project for exposing a local SQLite database and a local LanceDB vector database with information on Chicago Public Schools
The server implements two tools:
query_schools_and_neighborhoods
: Excecute a SELECT query on a table of Chicago public schools and their neighborhoods called "schooltoneighborhood" with the following schema: (id INTEGER NOT NULL, created_at DATETIME NOT NULL, school_id INTEGER NOT NULL, school_name VARCHAR NOT NULL, neighborhood VARCHAR NOT NULL, PRIMARY KEY (id)).
query
query_school_websites
: Query a database of Chicago public school websites for context relevant to answering a given question.
question
and optional parameter school_name
.Get the SQlite database and the LanceDB vector database from the [cps-childcare] project(https://github.com/mdagost/cps-childcare).
On MacOS: ~/Library/Application\ Support/Claude/claude_desktop_config.json
On Windows: %APPDATA%/Claude/claude_desktop_config.json
Development/Unpublished Servers Configuration:
"mcpServers": {
"mcp-cps-data": {
"command": "uv",
"args": [
"--directory",
"/path/to/mcp-cps-data",
"run",
"mcp-cps-data",
"--sqlite-path",
"/path/to/cps_crawler.db",
"--lancedb-path",
"/path/to/embeddings.lancedb"
]
}
}
Since MCP servers run over stdio, debugging can be challenging. For the best debugging experience, we strongly recommend using the MCP Inspector.
You can launch the MCP Inspector via npm
with this command:
npx @modelcontextprotocol/inspector uv --directory /path/to/mcp-cps-data run mcp-cps-data --sqlite-path /path/to/cps_crawler.db --lancedb-path /path/to/embeddings.lancedb
Upon launching, the Inspector will display a URL that you can access in your browser to begin debugging.
{
"mcpServers": {
"mcp-cps-data": {
"env": {},
"args": [
"--directory",
"/path/to/mcp-cps-data",
"run",
"mcp-cps-data",
"--sqlite-path",
"/path/to/cps_crawler.db",
"--lancedb-path",
"/path/to/embeddings.lancedb"
],
"command": "uv"
}
}
}
Seamless access to top MCP servers powering the future of AI integration.