README.md

metadata

title: NegoptimAi
emoji: ⚡
colorFrom: indigo
colorTo: gray
sdk: docker
app_port: 7860
pinned: false

Negoptim AI

Bilingual (EN/FR) RAG chatbot for Users Love IT (ULiT) — answers questions about the NegOptim platform, commercial negotiations, procurement, promotions and assortment, renders rich dashboard cards for common questions, and can compose & send emails (demo requests, contact, follow-ups) with explicit user confirmation.

Layer	Stack
Frontend	Next.js 15 · TypeScript · Tailwind (deployed on Vercel)
Backend	FastAPI · ChromaDB · sentence-transformers e5-small (Docker on HF Spaces)
LLM	Groq free tier, with optional Cerebras/Gemini fallback — no paid LLMs

📖 Documentation

• ARCHITECTURE.md — system design, services, data flows
• DEPLOYMENT.md — local development + production deployment & redeploy workflow

---

Quick start (local)

# 1. Backend deps
cd backend
pip install -r requirements.txt

# 2. Configure — backend/.env needs at least:
#    GROQ_API_KEY=gsk_...        (free key: https://console.groq.com/keys)

# 3. Build the vector index (first run, and after editing knowledge/)
python ../scripts/ingest.py

# 4. Run backend
uvicorn app.main:app --reload --port 8000

# 5. Frontend (second terminal)
cd frontend
npm install
npm run dev

Open http://localhost:3000. API docs at http://localhost:8000/docs.

Useful commands

Command (from backend/)	Purpose
python ../scripts/ingest.py [folder]	(Re)build the vector index
python ../scripts/test_email.py	Verify SMTP config (simulation or real)

Email runs in simulation mode (logged, not delivered) until SMTP_USER/SMTP_PASSWORD are set — see DEPLOYMENT.md §4 for all variables.