Wired n8n to Ollama and the MCP Tool — and Hit Almost Every Wall on the Way
n8n running as a container is a decent stand-in for "how would a real automation platform actually call this stuff". Everything up to now has been me calling things directly. Two integrations, same workflow: Ollama first, since it's simpler, then the MCP tool from Entry 08. Ollama went cleanly. Container running, one HTTP Request node pointed at http://host.docker.internal:11434/api/generate —…
The author describes the challenges faced while integrating n8n, Ollama, and the MCP tool in a complex automation workflow. Initially, n8n, running as a container, successfully called the Ollama integration. However, the MCP tool presented several issues, including a container running on a different filesystem and operating system than the host machine.
After switching from STDIO to SSE (Server-Sent Events) transport, the n8n container could communicate with the MCP server without any cross-filesystem or cross-OS compatibility problems. The author also faced a TypeError when trying to run the FastMCP server, which was resolved by correctly configuring the host and port parameters.
The network connection issue between n8n and the MCP server was due to a mislabeled endpoint field, which was corrected after testing the Podman hostname resolution within the container. The author emphasizes that the integration of multiple systems is not just a sum of their individual failures, but a compounded issue that requires careful attention to each layer to ensure successful integration.
The key takeaway from the series is that the underlying system remains consistent across different paths, producing the same output, which confirms the correct wiring and functionality of the entire setup.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
