AI COO — Multi-Agent Autonomous Operations System

A multi-agent system that runs your startup's operational functions autonomously using Claude AI.

Agents

Agent	Domain	Triggers
Dev Activity	GitHub PRs, CI, issues	Every 30m + webhooks
Outreach	Cold email campaigns, reply tracking	Every 2h
Marketing	Reddit, X, LinkedIn content	Daily
Finance	MRR, runway, expenses	Daily
PM	Sprint health, blockers, velocity	Every 6h
Research	Competitors, trends, leads	Every 12h
Legal	Contracts, compliance signals	Daily

Architecture

Next.js Frontend
      │  REST + Supabase Realtime
      ▼
FastAPI Backend (this repo)
      │
      ├─ core/context.py     Global context read/write
      ├─ core/events.py      Event bus (emit + consume)
      ├─ core/approvals.py   Human-in-the-loop approval queue
      └─ core/notifications.py  Push notifications
      │
      ▼
Celery Workers (one task per agent)
      │
      ▼
Supabase (PostgreSQL + Realtime + Storage)
Redis (Celery broker + result backend)

Quick Start

1. Install dependencies

cd backend
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt

2. Configure environment

cp .env.example .env
# Edit .env with your real credentials

3. Apply database migrations

Run in the Supabase SQL editor:

1. supabase/migrations/001_core_tables.sql
2. supabase/migrations/002_agent_domain_tables.sql

4. Seed global context

python ../scripts/seed_global_context.py

5. Start the API server

uvicorn app.main:app --reload

6. Start Celery worker + beat

# Worker (in a separate terminal)
celery -A celery_app worker --loglevel=info

# Beat scheduler (in a separate terminal)
celery -A celery_app beat --loglevel=info

7. Test the event cascade

python ../scripts/test_cascade.py

API Endpoints

Endpoint	Method	Description
/api/context/global	GET	Read global business context
/api/context/global	PATCH	Update global context fields
/api/events	GET	List recent events
/api/events/emit	POST	Emit a test event
/api/approvals	GET	List pending approvals
/api/approvals/{id}/respond	POST	Approve or reject
/api/notifications	GET	List notifications
/api/{agent}/run	POST	Manually trigger an agent
/api/{agent}/status	GET	Agent run status
/health	GET	Health check

Project Structure

See docs/AGENT_GUIDE.md for how to build or extend agents.