[SOLVED] Connection Refused ([Errno 111]) When Using Ollama with AutoGPT

Connection Refused ([Errno 111]) When Using Ollama with AutoGPT

I'm following this tutorial to integrate Ollama (http://localhost:11434/) with AutoGPT (http://localhost:3000) using Docker Compose. However, I encounter the following error when trying to connect:

Error calling LLM: [Errno 111] Connection refused

I ensured that:

Ollama is running.

$ curl http://localhost:11434/
Ollama is running

The host and port are correct (Ollama defaults to port 11434).

Error Details:

Connection Refused: Make sure Ollama is running and the host address is correct (also make sure the port is correct, its default is 11434).

Solution

Solution:
Instead of using localhost, replace it with host.docker.internal in the connection string:

http://host.docker.internal:11434/

Background:
Docker containers have their own network namespace, so localhost inside a container refers to the container itself—not the host. As a result, services running on the host aren't accessible via localhost from within the container.

Reference:
This workaround is discussed in this GitHub issue: Ollama Issue #703.