AJ.AI

Your own orchestration AI — chat, code/app generation, and image generation, unified behind one API so end users never touch a key.

What's here (v1)

• POST /api/chat — streaming conversational endpoint (Groq LLaMA 3.3 70B)
• POST /api/build-app — give it a plain-English app idea, get back generated file structure + code (React + optional Express backend)
• POST /api/generate-image — free, keyless image generation (Pollinations.ai)

Setup

1. Upload this whole folder to a new GitHub repo (drag-and-drop works fine on GitHub's web upload — no terminal needed).
2. Get a free Groq API key: https://console.groq.com
3. Deploy backend/ to Render (free tier):
   • New Web Service → connect your repo → root directory backend
   • Build command: npm install
   • Start command: npm start
   • Add environment variable GROQ_API_KEY in Render's dashboard (never commit your real .env — only .env.example is in this repo)
4. Once deployed, Render gives you a live URL like https://aj-ai.onrender.com — that's your API, callable from anywhere.

Roadmap

• Phase 1: chat + code generation
• Phase 1: image generation
• Phase 2: n8n workflow automation (self-hosted) to chain generate → test → auto-push to GitHub
• Phase 3: video generation (evaluating free-tier backends)
• Phase 4: frontend UI + auth + persistent memory

Example: build an app

POST /api/build-app
{ "prompt": "build a Duolingo-style language learning app with streaks and XP" }

Returns JSON with generated file paths + full code content for each file.

Notes

• No API key is exposed to end users of AJ.AI — only you (the owner) configure keys once in Render's environment variables.
• llama-3.3-70b-versatile is free on Groq with generous rate limits — good enough for both chat and code generation.
• Image gen requires zero setup — Pollinations.ai needs no key at all.