Senzing MCP Server

# Setting Up Senzing MCP Server with ChatGPT Desktop ## Prerequisites - ChatGPT Desktop App installed (with MCP support) - Senzing MCP server cloned (adjust paths below to match your installation location) ## Installation Steps ### 1. Locate ChatGPT Configuration Directory The ChatGPT Desktop app stores MCP configuration in: **macOS:** ``` ~/Library/Application Support/ChatGPT/ ``` **Windows:** ``` %APPDATA%\ChatGPT\ ``` **Linux:** ``` ~/.config/ChatGPT/ ``` ### 2. Create MCP Configuration Create or edit the file `mcp_config.json` in the ChatGPT configuration directory: **macOS:** ```bash # Create the directory if it doesn't exist mkdir -p ~/Library/Application\ Support/ChatGPT/ # Create the configuration file nano ~/Library/Application\ Support/ChatGPT/mcp_config.json ``` Add the following configuration (adjust paths to match your installation): ```json { "mcpServers": { "senzing": { "command": "/path/to/senzing-mcp-server/launch_senzing_mcp.sh", "env": { "SENZING_ENGINE_CONFIGURATION_JSON": "{\"PIPELINE\":{\"CONFIGPATH\":\"/etc/opt/senzing\",\"RESOURCEPATH\":\"/opt/senzing/g2/resources\",\"SUPPORTPATH\":\"/opt/senzing/data\"},\"SQL\":{\"CONNECTION\":\"sqlite3://na:na@/var/opt/senzing/sqlite/G2C.db\"}}", "LD_LIBRARY_PATH": "/opt/senzing/lib", "PYTHONPATH": "/opt/senzing/sdk/python" } } } } ``` **Configuration Notes:** - Replace `/path/to/senzing-mcp-server/` with the actual path where you cloned the repository - Update `SENZING_ENGINE_CONFIGURATION_JSON` with your database connection details - Update `LD_LIBRARY_PATH` to point to your Senzing library directory - Update `PYTHONPATH` to point to your Senzing Python SDK location **Note**: The launch script runs directly from source without requiring pip installation. All environment variables are set in the MCP configuration above. ### 3. Restart ChatGPT Desktop Close and restart the ChatGPT Desktop application to load the new MCP server. ## Verification After restarting ChatGPT, you should see the Senzing MCP server available. To test: 1. Start a new conversation in ChatGPT 2. Ask: "What MCP servers are available?" 3. You should see "senzing" listed 4. Try a query: "Search for entities named Robert Smith" ## Available Tools The Senzing MCP server provides 7 read-only tools: 1. **search** - Search for entities by attributes (name, address, phone, email, DOB) 2. **get_entity** - Get comprehensive details for a specific entity by entity ID 3. **get_source_record** - Look up entity by source record ID (e.g., CUSTOMERS:1001) 4. **find_relationship_path** - Find the connection path between two entities 5. **expand_entity_network** - Expand network of related entities to (n) degrees (max 3) 6. **explain_why_entities_related** - Explain why two entities are related (WHY analysis) 7. **how_entity_resolved** - Explain how an entity was resolved from records (HOW analysis) ## Example Queries - "Search for all entities with the name John Smith" - "Get the full details for entity ID 1" - "Show me how entity 1 was resolved" - "Find the relationship path between entity 1 and entity 8" - "Explain why entity 1 and entity 8 are related" - "Find the network of entities related to entity 1" ## Response Formatting for HOW/WHY Analysis For better formatted explanations of entity resolution, you can provide ChatGPT with formatting instructions: **Option 1: Add to Custom Instructions** 1. Go to ChatGPT Settings → Personalization → Custom Instructions 2. Copy the contents of `RESPONSE_FORMATTING.md` from the repository 3. Paste into "How would you like ChatGPT to respond?" 4. ChatGPT will automatically format all Senzing HOW/WHY results **Option 2: Include in Conversation** When asking about entity resolution, add: ``` Please format the results according to RESPONSE_FORMATTING.md ``` Then paste the contents of the formatting guide or upload the file. **Result:** Instead of raw JSON, you'll get clear summaries with step-by-step breakdowns, highlighted confirmations (✅) and denials (❌), and side-by-side comparisons. See the main README's "Response Formatting Guide" section for more details and examples. ## Troubleshooting ### MCP Server Not Appearing 1. Check that `mcp_config.json` is in the correct directory 2. Verify the launch script is executable: `ls -la launch_senzing_mcp.sh` 3. Check ChatGPT logs for errors ### Connection Errors 1. Test the launch script manually with environment variables: ```bash LD_LIBRARY_PATH=/opt/senzing/lib \ PYTHONPATH=/opt/senzing/sdk/python \ SENZING_ENGINE_CONFIGURATION_JSON='{"PIPELINE":...}' \ /path/to/senzing-mcp-server/launch_senzing_mcp.sh ``` 2. Verify environment variables are correctly set in your MCP config: ```bash cat ~/Library/Application\ Support/ChatGPT/mcp_config.json | grep -A 5 '"env"' ``` ### No Data Returned 1. Verify the Senzing database has data 2. Check that your `SENZING_ENGINE_CONFIGURATION_JSON` in the MCP config points to the correct database 3. Review logs in ChatGPT application ## Notes - This is a **read-only** MCP server - no data modification operations are available - The server starts fresh for each ChatGPT session - All queries go through the configured Senzing repository - Environment variables are passed from the MCP configuration to the server