Omo (oh-my-opencode)

AI coding agent container. Runs oh-my-opencode with local LLM inference via Bifrost.

Location

agap_git/omo/

cd ~/agap_git/omo
docker compose up -d

Exec into the container and run the agent against a project:

docker exec -it omo sh
cd /workspace/my-project
oh-my-opencode run "your task here"

opencode.json — mounted at /root/.config/opencode/opencode.json
Provider: Bifrost (http://bifrost:8080/v1) — OpenAI-compatible gateway to local Ollama models
Default model: ollama/qwen3:8b (GPU)
Available models: qwen3:8b, qwen3:4b, qwen2.5:1.5b, gemma3:4b

Connected to adolf_default network to reach the bifrost container.

Host path	Container path	Purpose
`/home/alvis`	`/workspace`	Home directory for coding projects
`./opencode.json`	`/root/.config/opencode/opencode.json`	Provider config (read-only)