Isolated environment for AI-assisted development using qwencode and Podman.
Every project gets its own container to preserve its dependencies without cluttering the host system.
Configuration and logs are stored locally within each project directory.
For default configuration you need:
podman(small manual tweaks required fordocker)- LLM server (default config is for OpenAI-compatible server on localhost:8080, e.g.
llama.cpp)
You have two options for handling qwencode configuration:
Option A: Project-based configuration (Default)
Creates a .qwen/settings.json in each project directory if no global config exists.
make buildOption B: Global container configuration
Bakes global-settings.json directly into the container image (/root/.qwen/settings.json). Project directories remain clean.
make build GLOBAL_CONFIG=1Install the script:
- Globally (requires sudo):
sudo make install
- Locally (for current user):
make install PREFIX=~/.local
Navigate to your project directory and run:
qwen-ctTo remove the script:
sudo make uninstallTo remove all project containers and the base image:
make clean