From your terminal, forward
Install CyxCode, connect a model, and open a project. Bring your own provider credentials or configure a local model.
Install
Run in a shell on macOS, Linux, or Windows with WSL / Git Bash.
curl -fsSL https://raw.githubusercontent.com/code3hr/cyxcode/HEAD/install | bash
Installs the latest release into ~/.opencode/bin. The compatibility path is intentional.
Download the Windows archive from GitHub Releases ↗, extract it, and run cyxcode.exe. Choose a build compatible with your machine.
Connect
Launch cyxcode in your project. Run /connect to add a provider, then /models to select a model.
Use your own credentials for a supported provider. Model access and usage charges depend on that provider and your account.
Configure an OpenAI-compatible local endpoint, such as Ollama, to run a model on your own hardware.
CyxCode's anonymous requests to OpenCode Zen have been rejected by the service. A fork of the client does not automatically receive access to its hosted free tier.
Connect Zen with your own API key to check your account's access. Authenticated access to Zen's free models has not been verified in CyxCode. Other providers may offer free models under their own limits; local models use your own hardware.
Provider setup and model testing ↗Build
Run /cyxinit inside CyxCode to create the project state structure. Then start with a small, concrete
task.
Read this project's README and explain its structure. Don't modify any files yet.
Save conventions with /remember. Review captured fixes with /learn-patterns.
Run /update inside CyxCode to check for a release. For older installs with a broken update URL, rerun
the installer above.
Clone the repository, install dependencies with Bun, and start the development runtime.
git clone https://github.com/code3hr/cyxcode.git
cd cyxcode
bun install
bun run dev