Skip to content

nurudeen19/quorum

Repository files navigation

Quorum

Pre-meeting intelligence for when there's no time to prepare.

That meeting invite just landed, and all you have is a name and a company. Quorum fills the gaps—pulling public context on people and organizations, turning minimal info into a clear pre-meeting memo so you can walk in informed instead of blindsided.

What it does

  • Takes basic meeting info (names, companies, what you need to accomplish)
  • Gathers public context about the people and organizations involved
  • Synthesizes it into a tight briefing memo you can skim in minutes
  • Lets you ask follow-up questions in chat if you need more detail
  • Streams progress in real-time so you see what it's researching

Get started in 5 minutes

Backend (API server)

cd backend
uv sync --extra transformer-cpu
# uv sync --extra transformer-gpu # selects the gpu variant of transformers required for prmpt guard
uv run scripts/migrate_db.py
uv run uvicorn app.main:app --reload

→ See Backend README for full setup and environment variables.

Frontend (Web UI)

cd frontend
npm install
npm run dev

→ See Frontend README for full setup and commands.

Docker (recommended for production & testing)

Run the entire stack (frontend + backend + PostgreSQL + Redis) with a single command:

# Start everything
docker compose up -d

# Backend API:    http://localhost:7860
# Frontend UI:    http://localhost
# API Docs:       http://localhost:7860/docs

Or start individual stacks independently:

# Backend + Postgres + Redis only
cd backend && docker compose up -d

# Frontend only (requires backend already running on quorum-network)
cd frontend && docker compose up -d

Environment setup:

  1. Copy the Docker env template: cp backend/.env.docker.example backend/.env
  2. Edit backend/.env and fill in your API keys (LLM providers, search, etc.)
  3. Run docker compose up -d from the root

Optional — Log streaming with Grafana:

Uncomment the loki and grafana services in backend/docker-compose.yml, then:

docker compose up -d
# Grafana: http://localhost:3000 (anonymous access, logs auto-discovered)

Note: The root Dockerfile (Dockerfile) is used by Hugging Face Spaces and requires no changes. The frontend uses a two-stage Node → Nginx build in frontend/Dockerfile.

How it's built

Layer Technology Purpose
Backend FastAPI, Python REST API, agent orchestration, database
Frontend Vue 3, TypeScript, Vite Interactive web interface, real-time updates
Agents LangGraph, LLMs Multi-step reasoning pipeline
Database PostgreSQL, SQLAlchemy Persist users, conversations, history
Observability LangSmith, OpenTelemetry Monitor and trace agent behavior

Next steps

About

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors