Skip to content

Makefile Reference

The Makefile is the recommended interface for Docker usage.

Target Purpose
make build Build Docker images.
make up Start the Ollama service.
make down Stop services.
make logs Follow Ollama logs.
make pull-model MODEL=... Pull a model inside the Ollama container.
make list-models List Ollama models.
make ingest DOCS_DIR=... MODEL=... Index reference documents.
make query QUERY_PATH=... EXT=... MODEL=... Analyze source code.
make shell Open the runtime app shell.
make dev-shell Open the dev container shell.
make format Run Ruff format in the dev container.
make test Run pytest in the dev container.

Query Variables

Variable Description
QUERY_PATH File or directory to analyze.
EXT File extension filter for directories.
MODEL Ollama model.
OLLAMA_URL Ollama API URL inside Docker.
NUM_CTX Ollama context window.
CHANGED_ONLY Set to 1 to analyze changed files only.
CHANGED_BASE Git base ref for changed-file analysis.
STAGED Set to 1 to analyze staged files only.
HOST_OLLAMA Set to 1 to use an Ollama running on the host instead of the compose service.

Using a host Ollama (HOST_OLLAMA=1)

If you already run Ollama natively on the host, starting the compose ollama service too fails with address already in use on port 11434. Set HOST_OLLAMA=1 to skip the compose service and talk to the host instead:

make query QUERY_PATH=/abs/path EXT=py CHANGED_ONLY=1 HOST_OLLAMA=1 MODEL=qwen2.5-coder:7b-instruct

This passes --no-deps and points OLLAMA_URL at http://host.docker.internal:11434 (reachable via the app service's extra_hosts entry). The host Ollama must listen on all interfaces, not just localhost — start it with OLLAMA_HOST=0.0.0.0:11434 (a native install bound to 127.0.0.1 will refuse the container's connection).