Lemonade

What is Lemonade?

Lemonade is a local AI server that provides cloud API–like capabilities for chat, coding, speech, and image generation, running 100% free and privately on the user’s own NPU and GPU. It comes in two forms: Lemonade Server (installable service that connects apps via OpenAI, Anthropic, and Ollama APIs) and Embeddable Lemonade (portable binary for packaging into applications). The project is built by the community with optimizations by AMD engineers.

How to use Lemonade?

Install Lemonade using the provided options for Windows, Linux, macOS, or Docker. After installation, browse and download models with the built-in Model Manager or the lemonade pull command. Then generate text, images, or speech using the built-in interfaces, or connect to hundreds of third‑party apps via standard APIs (e.g., lemonade launch claude for coding with Claude Code).

Key features of Lemonade

• Free, private local AI server (no cloud costs)
• Supports chat, coding, speech, and image generation
• Compatible with OpenAI, Anthropic, and Ollama API standards
• Runs on CPU, GPU, and NPU across multiple platforms
• Built-in Model Manager for browsing and downloading models
• Available as installable server or embeddable binary

Use cases of Lemonade

• Run LLMs locally for private chat and coding assistance
• Generate images with SDXL‑Turbo without cloud dependencies
• Perform speech synthesis and transcription (Whisper) locally
• Embed local AI into custom applications via Embeddable Lemonade
• Connect local AI to hundreds of apps like Open WebUI and Claude Code

FAQ from Lemonade

What models does Lemonade support?

Lemonade supports LLMs in GGUF, FLM, and ONNX formats, plus whisper, stable diffusion, and other models across CPU, GPU, and NPU. Users can also import custom GGUF/ONNX models from Hugging Face.

What hardware is required?

Lemonade runs on Apple Silicon GPU (Metal), NVIDIA GPUs (Turing or newer, CUDA), AMD GPUs (ROCm, Vul