Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@kg-mcpsearch for 'machine learning' and show its connections"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
kg-mcp
Knowledge Graph MCP Server. Maschinenoptimiert – kein Prosa, kein Markdown. Kompaktes JSON für LLM-Kontexteffizienz.
Deployment (Coolify)
GitHub Repo erstellen und Code pushen
In Coolify: Neue Application → GitHub → dieses Repo
Build Pack: Dockerfile
Port: 8099
Domain:
kg-mcp.kailohmann.deVolume:
/data(persistent für SQLite)Keine weiteren Env-Vars nötig (Defaults sind gesetzt)
Claude MCP Integration
Nach Deployment in Claude.ai unter Settings → MCP Servers hinzufügen:
URL: https://kg-mcp.kailohmann.de/mcp
Name: kg-mcpTools
Tool | Funktion |
| Session-Start: State + Node-Index |
| Node + N-Hop Nachbarschaft |
| Volltext-Suche |
| Node erstellen/aktualisieren |
| Edge erstellen/aktualisieren |
| Node löschen |
| Edge löschen |
| Projektstatus setzen |
| Batch-Operationen |
| BFS Subgraph-Traversierung |
Compact Format
Node: {"id":"schwamm","t":"metaphor","s":"...","b":[1,3,6],"st":"explored","k":"...","m":{}}
Edge: {"src":"schwamm","rel":"contrasts","tgt":"basic_programm","w":1.0,"n":"..."}
Index: {"id":"schwamm","t":"metaphor","st":"explored"}Lokal testen
KG_DB_PATH=./test.db KG_TRANSPORT=streamable-http python server.pyThis server cannot be installed
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.