publicai

Public AI MCP Server provides access to community-maintained information via Wiki Tools and real-time external APIs. It exposes functions that read data from the wiki.publicai.co cargo-backed tools (for crisis resources, community events, and other local data) and also calls external APIs for live transit, parking, and location services. AI assistants can invoke these MCP functions to fetch fresh information, then present it to users with up-to-date resources such as crisis hotlines, transit departures, or carpark availability. The server is designed to be queried by an AI agent, returning structured data that the assistant can format into helpful responses.

Core capabilities include: wiki-based read operations (list_tools_by_community, use_tool) to retrieve community-maintained resources; and API-based functions for real-time data such as Swiss transport (search_swiss_stations, get_swiss_departures, plan_swiss_journey), Singapore carpark availability, and OpenStreetMap lookups (search_osm_nominatim). These tools enable use cases like crisis support, transit planning, and locating local resources, with clear examples provided in the repository’s docs. To use, configure the MCP server in your client (e.g., Claude Desktop) and call the available functions as shown in the examples to obtain current data and workflows.

Prerequisites:

• Python 3.8+
• FastMCP

Setup:

1. Clone the repository:

git clone https://github.com/yourusername/pai-mcp-server.git
cd pai-mcp-server

2. Install dependencies:

pip install -r requirements.txt

3. Run the server:

python main.py

The server will start on http://127.0.0.1:8000 (default).

Configuration (example for Claude Desktop):

• Add to your Claude Desktop config (macOS path shown as example):

{
  "mcpServers": {
    "publicai": {
      "command": "python",
      "args": ["/path/to/pai-mcp-server/main.py"]
    }
  }
}

Environment variables: Currently, no environment variables are required. Future API integrations may require API keys (these should be added as env vars and documented in the function code). If you modify functions, ensure proper error handling and timeouts (the examples suggest a 10s default). When running locally, ensure port 8000 is accessible to clients and that any external API keys are securely managed. If data from the Wiki Tools changes, you may need to refresh or re-index the Cargo data on the wiki server to keep responses current.