Lychee
A local LLM runtime and orchestration layer, positioned as a drop-in Ollama replacement, that pulls models from Hugging Face, serves them, and offers a built-in dashboard, pipelines and structured output.
Lychee is a local runtime for pulling, running and orchestrating large language models. A single command starts a server, and another pulls a Hugging Face GGUF model by its repository name and starts chatting with it; the pull-and-run mode starts the chat as soon as the model is ready. A browser-based dashboard at a local address manages models, monitors performance and offers interactive chat.
It installs from source with Go 1.22 or later, or through one-line install scripts for Windows, macOS and Linux that try a Go install first and fall back to a prebuilt release binary, with Docker images also available. The project describes 24 CLI commands covering model management, benchmarking and pipelines, and is written in Go.