ProxyGateLLM v1.0.0 - Multi-LLM Gateway

🚀 ProxyGateLLM v1.0.0 - Multi-LLM Gateway

A free multi-LLM gateway supporting Gemini, OpenAI, Claude, and Ollama with priority fallback and SHA256 caching.

Features

• 🤖 Multi-provider LLM support (Gemini, OpenAI, Claude, Ollama)
• 🔄 Priority-based fallback system
• 💾 SHA256 response caching
• 🌐 RESTful API with Express.js + TypeScript
• 📊 Request logging and monitoring

Looking for Contributors! 🤝

We welcome contributions of all sizes:

• Additional LLM provider integrations
• Streaming response improvements
• UI dashboard development
• Testing and documentation
• Bug fixes and performance improvements

Check our good first issues to get started!

JSUPTER AI Gateway v2.1.0

v2.1.0 — Code Audit & Infrastructure

Improved

• Code audit — verified all provider adapters (Z.ai, Qwen, Puter.js)
• Verified all API endpoints
• Verified task classifier, routing engine, streaming, middleware
• Added Dockerfile for containerized deployment
• Added lint and clean scripts
• Improved test script with graceful fallback

Endpoints

• POST /chat — Auto-routed unified chat
• POST /v1/chat/completions — OpenAI-compatible
• POST /v1/messages — Anthropic-compatible
• POST /zai/chat — Direct Z.ai route
• POST /qwen/chat — Direct Qwen route
• POST /route — Routing decision only
• GET /health — Health check
• GET /models — Available models
• GET /status — Provider status

Providers

• Z.ai — reasoning, general intelligence
• Qwen 2.5 Coder 32B — code, structured output
• Puter.js — 18+ models (GPT-4o, Claude, DeepSeek, etc.)

v2.0.0 — JSUPTER AI Gateway: Multi-LLM Task-Based Router

JSUPTER AI Gateway v2.0.0

Breaking Change

Complete architecture overhaul from simple Puter.js proxy to Multi-LLM Gateway System with task-based routing, streaming, and multi-provider abstraction.

New: Multi-LLM Gateway System

• Task Classification Engine — classifies input as code/reasoning/infra/multimodal/structured
• Task-Based Routing — auto-routes to best provider with fallback chains
• Provider Abstraction — Z.ai (reasoning), Qwen (code), Puter.js (18+ models)
• SSE Streaming — token-by-token streaming via Server-Sent Events
• Hybrid Execution — Qwen generates base, Z.ai reviews (for infra tasks)

New: 9 API Endpoints

Endpoint	Purpose
POST /chat	Unified auto-routing interface
POST /v1/chat/completions	OpenAI-compatible (streaming)
POST /v1/messages	Anthropic-compatible
POST /zai/chat	Direct Z.ai route
POST /qwen/chat	Direct Qwen route
POST /route	Routing decision only
GET /health	Health check
GET /models	List available models
GET /status	Provider status

New: Community & Docs

• Trilingual README (English, Bahasa Indonesia, 中文)
• MIT License
• CONTRIBUTING.md, CODE_OF_CONDUCT.md, SECURITY.md
• Issue Templates + PR Template

---

Full Changelog: v1.0.0...v2.0.0