Lychee v0.1.1

Cross-platform binaries

🍒 Lychee v0.1.0-alpha — First Public Release

🍒 Lychee v0.1.0-alpha

What is Lychee?

Lychee is a drop-in upgrade for Ollama — a local LLM runtime with enterprise-grade features built on top of the battle-tested Ollama inference engine.

✨ What's New in This Release

🔌 Multi-Protocol API Gateway

• Anthropic Messages API (/v1/messages) — run Claude-compatible workloads locally with any model
• OpenAI Responses API (/v1/responses) — structured JSON output with full JSON Schema validation
• OpenAI Chat Completions (/v1/chat/completions) — drop-in OpenAI replacement

🤖 Model Composer (/api/compose)

Chain multiple local models together in a pipeline:

• Sequential execution with template substitution ({{input}}, {{step[n].output}})
• Parallel branch execution — run independent steps concurrently
• Per-step timeouts — prevent hanging pipelines
• Fallback model — automatically retry with a backup model on failure
• SSE streaming — real-time token streaming for each step

🛡️ Stateful Agent Sandbox

• Terminal UI approval gates for every tool call
• Auto-allow safe commands, deny dangerous patterns (fork bombs, privilege escalation)
• Path traversal prevention, hierarchical allowlisting

🌐 Embedded Web Dashboard (/dashboard/)

• Full React SPA served directly from the binary
• Chat interface, model management, system resource monitoring

📥 Direct HuggingFace Pull (\lychee hf pull)

• Download GGUF models directly from HuggingFace Hub
• Multi-shard concurrent downloads with resume support
• SHA256 integrity verification

📊 Developer Tools

• \lychee compare\ — benchmark multiple models side-by-side
• \lychee stats --tui\ — live terminal performance dashboard
• \lychee scan\ — hardware detection and model recommendations
• \lychee generate-client\ — generate SDK boilerplate in Python/JS/Go/Rust

🚀 Quick Start

\\�ash
git clone https://github.com/MD-Mushfiqur123/lychee.git
cd lychee
go build -o lychee .
./lychee serve
\\

📋 Requirements

• Go 1.22+
• A GPU (NVIDIA/AMD/Apple Silicon) or CPU-only mode

⚠️ Alpha Notice

This is an alpha release. Core features are functional but not yet production-hardened. Use with care in production environments.

---

Built on top of Ollama — the open-source local LLM runner.

Lychee v0.3.0-alpha - SDKs and DAG Composer

Python SDK, JavaScript SDK, DAG conditional branching, SkipOnError resilience added.

Lychee v0.2.0 Alpha

What's New in v0.2.0-alpha

🚀 New Features

• Universal Pull: Auto-detect HuggingFace models with lychee pull org/model
• --run flag: Pull + create + chat in one command
• Interactive REPL: lychee interactive with /model, /system, /clear commands
• Built-in Dashboard: Browser UI at localhost:11434
• Shell Completions: bash, zsh, fish, PowerShell
• Self-update: lychee update command
• Benchmark: Compare model performance

🖥️ Desktop App

• Lychee Desktop v0.1.0 — native GUI (Chat + Studio + Model Manager)
• Auto-start lychee serve from desktop
• Pipeline export/import/share

📦 SDKs

• Python SDK: pip install lychee
• JavaScript SDK: npm install lychee-js
• Go SDK: go get github.com/MD-Mushfiqur123/lychee/sdk/go
• Ruby SDK: gem install lychee

🐳 Docker

• Multi-arch: linux/amd64 + linux/arm64
• Healthcheck + compose support

🔧 Fixes

• 6 bugs fixed including critical runtime panic in generate command
• Better error messages
• Config file support (~/.lychee/config.yaml)

Full changelog: https://github.com/MD-Mushfiqur123/lychee/blob/main/CHANGELOG.md