AI Potluck
Contribute
Model components / Inference code

Lemonade

Lemonade SDK

Lemonade is a local AI server that runs chat, coding, speech and image models on the user's own NPU and GPU, exposing OpenAI-, Anthropic- and Ollama-compatible APIs that hundreds of apps can connect to. It installs on Windows, macOS and Linux with GGUF and ONNX backends, and an embeddable binary packages the same runtime into other applications. The community builds it with optimizations from AMD engineers for Ryzen AI, Radeon and Strix Halo PCs.

The PyPI package lemonade-sdk named by the census no longer exists (404 this session), so no package artifact is declared and adoption reads the stars route. Verified 2026-09-01 via GitHub, the LICENSE body, the repository README and the PyPI JSON endpoint's 404.

Openness

5 high confidence
5.0
license
Apache-2.0(OSI
source
public
core-gated
ungated(free local server

The LICENSE body carries the full unmodified Apache-2.0 text ("Copyright 2026 Lemonade Community" in the appendix slot), followed after END OF TERMS AND CONDITIONS by attribution blocks for vendored components - Lucide icons, amdxdna_accel.h, llama.cpp, Ollama, TurnkeyML (Groq copyright), aixlog - each under its own permissive license. Those are notices about bundled third-party code, not conditions on Lemonade, so the governing license reads as clean Apache-2.0. The product is entirely free and self-hosted; nothing is withheld.

  • https://raw.githubusercontent.com/lemonade-sdk/lemonade/main/LICENSE recorded 2026-09-01

    LICENSE body, read in full: unmodified Apache-2.0 terms ending at END OF TERMS AND CONDITIONS, "Copyright 2026 Lemonade Community", then vendored-component attribution sections (Lucide, amdxdna_accel.h, llama.cpp, Ollama, TurnkeyML, aixlog), each MIT-family, with no added condition on the product itself.

  • https://api.github.com/repos/lemonade-sdk/lemonade recorded 2026-09-01

    Repo metadata for lemonade-sdk/lemonade - archived false, license Apache-2.0, 5,560 stars, homepage https://lemonade-server.ai/, pushed 2026-09-01.

  • https://raw.githubusercontent.com/lemonade-sdk/lemonade/main/README.md recorded 2026-09-01

    README, read for a paid tier (none - "100% free and private"): local AI server with OpenAI/Anthropic/Ollama APIs, Lemonade Server and Embeddable Lemonade flavors, Windows/macOS/Linux/Docker/Snap/PPA distribution, chat, image gen, speech gen and transcription models on NPU and GPU.

Adoption

2 low confidence
2.0

5,560 GitHub stars, in the 1K-10K band of the stars scale. The product ships as OS installers, Docker images, a Snap and a PPA - channels the pipeline cannot count - and the PyPI package the census named has been removed (404 this session), so stars are the only measurable route.

Capability

3 medium confidence
3.0

Two bands below the vllm anchor, level with llamafile and uzu among the local engines. Its NPU coverage and multi-modality are broader than the other single-node locals, but the serving engines underneath are llama.cpp and ONNX Runtime and it carries no distributed or datacenter surface.

Verified 2026-09-01