Lemonade

Lemonade is a local AI runtime that lets a computer run text, image and speech models on its own hardware. Installation provides an AI runtime, a graphical interface, a command line tool and API endpoints, so a user can chat with models directly, connect them to agents, or build them into an application. It is aimed at developers and enthusiasts who want private, on device AI. The vendor describes it as omni modal, covering chat, vision, image generation, speech, transcription and embeddings, and states that it collects zero telemetry. Because it follows the OpenAI API standard, many existing apps work with it out of the box, and the project also offers an embeddable SDK for teams that want to ship a copy of Lemonade inside their own software. Installers are listed for Windows, macOS, Docker, Snap, Ubuntu, Arch, Fedora and Debian. It is local first, and can route requests to other servers and cloud APIs on demand. Lemonade is free to use. The vendor publishes no paid plans, and its home page states there is no upsell. The project is developed in the open with a public GitHub repository, working groups that steer the roadmap, and a Discord community, and it ships a new release every week.