Lemonade

lastmod 2026-04-03created 2026-04-031 edits

Fast open-source local LLM server using GPUs and NPUs

About

• An AMD-backed local AI server: an open-source platform that uses GPUs and NPUs to process text, images, and speech quickly
• Emphasizes local execution and privacy, and is compatible with the OpenAI API standard, so it integrates with many apps right away
• Provides a practical local AI environment with a lightweight C++ backend, automatic hardware configuration, and concurrent multi-model execution
• Supports Chat, Vision, Image Generation, Transcription, Speech Generation through a single unified API
• Provides the same environment on Windows, Linux, macOS(beta), and its built-in GUI makes downloading and switching models easy

See also

• LLM
• GPU
• NPU
• OpenAI API
• llama.cpp
• Ryzen AI
• AMD

Favorite site

• Lemonade official site
• GitHub repository
• Lemonade by AMD: Fast open-source local LLM server using GPUs and NPUs - GeekNews