Menu Documentation

Install

One binary on macOS, Linux or Windows — or Docker, or Ollama.

Ogcode ships as a single static binary. Pick your platform below; every path ends with the same thing — an ogcode command on your PATH.

The install script downloads the latest release for your architecture and drops the binary in /usr/local/bin:

curl -fsSL https://ogcode.xyz/install.sh | sh

Or use Homebrew:

brew tap prasenjeet-symon/tap
brew install ogcode

Point it at a model

Ogcode needs one LLM credential. It reads these from the environment, and you need at least one:

VariableProvider
ANTHROPIC_API_KEYAnthropic / Claude
OPENAI_API_KEYOpenAI / GPT
OPENROUTER_API_KEYOpenRouter
OLLAMA_BASE_URLOllama or any OpenAI-compatible local endpoint
export ANTHROPIC_API_KEY=sk-ant-...

You can also keep keys in ~/.config/ogcode/config.json for global settings, or ogcode.json at a project root for project settings. Environment variables win over both.

For a local model with no API key at all:

ollama serve
ogcode

Verify it

ogcode version      # prints the version and commit
ogcode check-updates

check-updates exits with a non-zero status when a newer release is available, which makes it usable in a script.

Building from source

If you would rather build it yourself, a source build needs the web UI compiled first. The Makefile does both:

make build          # builds web/dist, then the Go binary with CGO_ENABLED=1
make install        # copies to ~/.local/bin and re-signs on macOS

CGO_ENABLED=1 is required because the binary links native libraries for PDF rendering and syntax parsing.

Next: Quick start.

From side project to enterprise codebase.

One Go binary, any model, your keys. Ogcode builds the context each task needs, so you get the job done without burning through your wallet.

Or install from a terminal: curl -fsSL https://ogcode.xyz/install.sh | sh