Air-gapped

Open WebUI Computer runs without internet access once installed. The Python wheel bundles the built frontend assets and the Docker image is self-contained, so nothing is fetched at runtime.

Build the artifacts on a connected machine

# Python route: download a wheelhouse
pip download --dest wheelhouse 'cptr[all]'

# Docker route: save the image
docker pull ghcr.io/open-webui/computer:latest
docker save ghcr.io/open-webui/computer:latest -o cptr-image.tar

Download the wheelhouse on a machine matching the offline host's platform and Python version so pip picks compatible wheels.

Install on the offline host

From the wheelhouse:

python -m venv .venv
. .venv/bin/activate
pip install --no-index --find-links ./wheelhouse 'cptr[all]'
cptr run --host 0.0.0.0

Or load and run the Docker image:

docker load -i cptr-image.tar
docker run --rm -it \
  --network=none \
  -p 8000:8000 \
  -v cptr-data:/data \
  -v "$PWD:/workspace" \
  -w /workspace \
  ghcr.io/open-webui/computer:latest

If pip still tries to contact an index, check that both --no-index and --find-links ./wheelhouse are present.

What still needs a network

Files, editor, terminal, git (local operations), and workspaces all work fully offline. Anything that talks to an external service does not:

• Hosted model APIs (OpenAI, Anthropic, and other cloud providers)
• Web search providers
• Git remotes (push, pull, clone from the internet)
• Messaging platforms (Telegram, Discord, Slack, WhatsApp, Signal)
• Remote MCP and OpenAPI tool servers

Local models still work: point the app at an Ollama or other OpenAI-compatible endpoint reachable on your own network.