Lemonade, created by AMD, is a server software plus GUI for operating native AI fashions, much like tasks like LM Studio (or, extra distantly, ComfyUI). What it lacks in configurability, it tries to make up for in broader integration with third-party apps that use normal APIs, and with assist for non-NVIDIA runtimes.
Lemonade works with quite a lot of runtimes and back-end engines. It helps AMD GPUs, Ryzen NPUs, Vulkan, and CPU execution (though not for all duties), together with the llamacpp, whispercpp, sd-cpp, kokoro, ryzenai-llm and flm again ends. Along with offering its personal set of APIs, Lemonade interoperates with a broad set of business requirements together with OpenAI, Ollama, Anthropic, and llama.cpp. Each GGUF and ONNX fashions are supported.