Chorus

Multi-LLM PWA — broadcast prompts to multiple Azure AI Foundry models simultaneously, streaming responses into per-model tabs.

Stack

• Next.js 16 (App Router, React 19, TypeScript)
• Vercel AI SDK (@ai-sdk/azure) for streaming
• Zustand for state management
• shadcn/ui + Tailwind CSS v4 for UI
• next-pwa for Progressive Web App support

Quick Start

# 1. Clone
git clone https://github.com/alvetm/chorus.git && cd chorus

# 2. Install
npm install

# 3. Configure environment
cp .env.local.example .env.local
# Edit .env.local with your Azure AI Foundry API keys

# 4. Run
npm run dev

Open http://localhost:3000.

Environment Variables

Variable	Description
FOUNDRY_AUSTRALIAEAST_NAME	Resource name for Australia East endpoint
FOUNDRY_AUSTRALIAEAST_API_KEY	API key for Australia East endpoint
FOUNDRY_EASTUS2_NAME	Resource name for East US 2 endpoint
FOUNDRY_EASTUS2_API_KEY	API key for East US 2 endpoint

API keys are server-side only — they never reach the browser.

Architecture

Browser (PWA)  →  /api/chat (Next.js route)  →  Azure AI Foundry
                  /api/deployments            →  Model Discovery

• Same-origin proxy — no CORS issues, no browser flags needed
• Concurrent streaming — all selected models stream in parallel
• Per-model tabs — switch between responses with status badges

Supported Models

gpt-4.1-mini · gpt-5.2-chat · o4-mini · DeepSeek-R1-0528 · DeepSeek-V3.1 · grok-4-fast-reasoning · Llama-4-Maverick