Install Veyyon.
One command, no build step. Veyyon is a single binary from GitHub releases. Install it, add a key, and open a repo.
curl -fsSL https://get.veyyon.dev | sh
Install
curl -fsSL https://get.veyyon.dev | sh — or grab a release binary from GitHub.
Sign in
Add a provider key with /login <provider>, set its env var, or point at any OpenAI-compatible endpoint.
Build
Run vey in a repo, pick your model with /model, and start.
macOS, Linux, Windows.
Prebuilt binaries per platform. The installer wires your PATH, shell completions, and a vey plugin doctor self-test.
macOS
Apple Silicon and Intel. curl … | sh or a signed release archive.
Linux
x64 and arm64. XDG-aware paths via vey config init-xdg.
Windows
In PowerShell: irm https://veyyon.dev/install.ps1 | iex — then the same vey command in your terminal.
Bring a key, or run local.
Any OpenAI-compatible endpoint works. Local engines are found automatically and need no key.
| Provider | How to sign in |
|---|---|
| Anthropic · OpenAI · Google · Groq · xAI | /login <id> · or set the provider env var |
| OpenRouter · Mistral · Bedrock | key or the AWS credential chain |
| Ollama · LM Studio · llama.cpp | auto-discovered, keyless — just run the engine |
Full install, XDG, and provider setup live in the handbook.