SMF Clearinghouse

What you'll get

A self-hosted ChatGPT-style web interface that talks to your local Ollama models and any OpenAI-compatible API. You can switch models, manage conversations, and expose the UI to your LAN if needed.

Prerequisites

Ollama running locally (see the Ollama CUDA recipe)
Docker or a local Python environment
At least 2 GB free disk space

Option 1: Docker (recommended)

docker run -d -p 3000:8080 \
  --add-host=host.docker.internal:host-gateway \
  -v open-webui:/app/backend/data \
  --name open-webui \
  --restart always \
  ghcr.io/open-webui/open-webui:main

Visit http://localhost:3000.

If Ollama is also in Docker, connect to it using the same network. If Ollama is on the host, use http://host.docker.internal:11434 as the Ollama API URL in Open WebUI settings.

Option 2: Local Python install

# Requires Python 3.11+
git clone https://github.com/open-webui/open-webui.git
cd open-webui
pip install -r requirements.txt -U
bash start.sh

Visit http://localhost:8080.

Step 1: Connect Ollama

The first time you open Open WebUI, it prompts for an Ollama endpoint:

http://host.docker.internal:11434

If running Open WebUI directly on the host, use:

http://localhost:11434

Click Save. The models you have pulled will appear in the model selector.

Step 2: Pull models through the UI

Open WebUI can pull Ollama models directly:

Settings → Models → Pull a model

Enter a model name such as qwen3.5:9b and wait for the download.

Step 3: Add an OpenAI-compatible API key

To compare local models against cloud models in the same interface:

Go to Settings → Connections.
Add an OpenAI API base URL and key, or an Anthropic/OpenRouter key.
The new provider appears in the model dropdown.

Step 4: LAN access (optional)

To expose Open WebUI to other devices on your network:

docker run -d -p 3000:8080 \
  -e ENV=prod \
  -e OLLAMA_BASE_URL=http://host.docker.internal:11434 \
  -v open-webui:/app/backend/data \
  --name open-webui \
  --restart always \
  ghcr.io/open-webui/open-webui:main

Then access it at http://your-lan-ip:3000.

Do not expose Open WebUI to the public internet without authentication.

Sanity check

Check	Command / Action
Container running	`docker ps`
Ollama reachable	`curl http://localhost:11434/api/tags`
WebUI logs	`docker logs open-webui`
Model list loads	Refresh the browser and check the dropdown

Common gotchas

"Cannot connect to Ollama" — If using Docker, verify --add-host=host.docker.internal:host-gateway is set and the Ollama endpoint is correct.
No models listed — Pull at least one model with ollama pull qwen3.5:9b.
Slow first load — Docker image download can take several minutes depending on connection speed.

Next step

Use Open WebUI as a chat playground to test prompts before using them with agents like Cline, Aider, or OpenClaw. Then move the best prompts into versioned prompt files for your agents.

Deploy Open WebUI with Ollama