Quick Start
Install Silk Code, connect a model, and point the coding agent at a repository.
1. Install or update
Requires Python 3.10 or newer and git. This zsh setup is safe to rerun and upgrades the packaging tools before installing Silk Code.
if [ -d "$HOME/SilkCode/.git" ]; then
git -C "$HOME/SilkCode" pull --ff-only
else
git clone https://github.com/RupertCloud/SilkCode.git "$HOME/SilkCode"
fi
cd "$HOME/SilkCode"
[ -d .venv ] || python3 -m venv .venv
.venv/bin/python -m pip install --upgrade pip setuptools wheel
.venv/bin/python -m pip install -e .
"$HOME/SilkCode/.venv/bin/silkcode" --help
grep -Fq 'export PATH="$HOME/SilkCode/.venv/bin:$PATH"' ~/.zshrc 2>/dev/null || \
echo 'export PATH="$HOME/SilkCode/.venv/bin:$PATH"' >> ~/.zshrc
source ~/.zshrc
silkcode --help
If an older environment reports that setup.py or setup.cfg is missing, the packaging-tool upgrade above is the fix. See Installation for troubleshooting.
2. Connect a model
# Cloud example
export DEEPSEEK_API_KEY=sk-your-key
# Or install a local model through Ollama
silkcode models pull qwen2.5-coder
Use a model with tool-calling support. Run silkcode models to inspect configured providers and models.
3. Open a project
cd ~/Projects/my-project
silkcode gui .
# Or pass the path directly
silkcode gui ~/Projects/my-project
silkcode ~/Projects/my-project
silkcode gui ~/SilkCode opens the Silk Code source repository itself. Pass another project directory to work on another codebase.
Next steps
Read the CLI reference, choose a model in Models, or learn how Silk Code handles sandboxed execution.