Ollama Integration

[GokulSoundararajan]

Could someone direct me how to do the same using ollama with Deepseek/Mistral

[itsuzef]

Ollama's API is REST-based rather than socket-based
You'll need to map MCP commands to appropriate Ollama API endpoints
The response format will need to be transformed to match your MCP protocol.

If you have any IDE that supports MCP you can do it! Let me know if you have any questions

[timglenn123]

try something like this. https://pypi.org/project/ollama-mcpo-adapter/
really impressed with the mcp server. i got some interesting things coming out of Claude sonnet with cursor. Although I'm running out of credits like crazy. So looking into how to run a local LLM. Docker just released their Docker Model Runner beta on 4.40.0 desktop.