ChatMock acts as a reverse proxy server that accepts OpenAI-compatible API requests and forwards them to OpenAI Codex, allowing users to access Codex through any OpenAI-compatible client or tool. The server runs locally and provides endpoints like /v1/chat/completions with model names like gpt-5.4 and gpt-5.5. It supports multiple installation methods including pip, Homebrew, Docker, and GUI installers, and can be integrated with tools like Ollama by using its base URL as the endpoint.