Lemonade-sdk adds image support to lemonade

By vramkickedin | | 2 min read
A digital pixelated lemon on a pixelated grass surface

Lemonade is an open-source local AI server that lets users run LLMs, and speech tools directly on their own hardware. The project provides a unified API that combines text, audio and now image generation capabilities under a single server interface. The developer aims to make local generative AI accessible without requiring cloud services.

Previously focused on text-based AI, the tool has expanded to include Stable Diffusion for image generation and Whisper for speech recognition. A one-click installer handles setup, giving users immediate access to models through consistent endpoints at localhost.

All-in-One Local AI Capabilities

  • One-click installer available for Windows, Linux, macOS (beta), and Docker.
  • Unified API endpoints for chat completions.
  • Built-in Model Manager to browse, download, and organize models.
  • Support for multiple backends including Vulkan, ROCm, CPU, and AMD NPU.
  • OpenAI-compatible API structure.
  • Desktop application with visual interface for testing models.

Anyone building AI-powered applications can use Lemonade to run multiple model types through a single server. Instead of managing separate installations for text, image, and speech tools, developers get one consistent API. The local execution keeps data private and eliminates recurring cloud costs.

Lemonade's Roadmap and Hardware Support

The development team has outlined clear plans for upcoming features. Future updates will add ROCm and Vulkan backends for AMD GPUs, along with NPU-specific builds for AMD's XDNA2 architecture. Text-to-speech capabilities and streaming audio input are also planned.

Current support includes various model formats such as GGUF, FLM, and ONNX across different hardware configurations. Users can import custom models directly from Hugging Face or select from the built-in library. As a community-driven project sponsored by AMD, Lemonade welcomes contributions through GitHub.

Lemonade simplifies running generative AI locally by combining multiple modalities into one unified server. The straightforward installation and consistent API design reduce complexity for practical AI deployment.

Get lemonade on GitHub or check out their project page