Deal Co-Pilot

A modern full-stack SaaS application for AI-powered investment due diligence research.

Web Application | FastAPI Backend | AI Agents | Beautiful UI

Quick Start - Full-Stack Application

# 1. Install Python dependencies
pip install -r requirements.txt

# 2. Install frontend dependencies
cd frontend && npm install && cd ..

# 3. Set API key
export OPENAI_API_KEY="your_key_here"

# 4. Run full-stack (backend + frontend)
./run_fullstack.sh

# 5. Open browser
open http://localhost:3000

That's it!

Or Run Separately

# Backend only (port 8000)
./run_server.sh

# Frontend only (port 3000)
cd frontend && npm run dev

What You Get

• Next.js 15 Frontend - Modern React with TypeScript
• Tailwind CSS - Beautiful, responsive design
• Natural Language Input - Just describe the company
• Real-time Progress - Watch research generate live
• Formatted Reports - Professional, cited research
• Fully Responsive - Works on all devices
• Fast API Backend - Python-based REST API
• Dual AI Agents - OpenAI & Gemini options

Quick Start - Command Line

Option 1: OpenAI (Simplest - No Tavily!)

# 1. Set up environment
export OPENAI_API_KEY="your_key_here"

# 2. Install dependencies
pip install -r requirements.txt

# 3. Run example
python -m deal_copilot.examples.example_run_openai

Option 2: Gemini + Tavily (Cost-effective)

# 1. Set up environment
export GOOGLE_API_KEY="your_gemini_key"
export TAVILY_API_KEY="your_tavily_key"

# 2. Install dependencies
pip install -r requirements.txt

# 3. Run example
python -m deal_copilot.examples.example_run

Project Structure

vinnie/
├── frontend/                         # Next.js Frontend
│   ├── app/
│   │   ├── globals.css              # Tailwind styles
│   │   ├── layout.tsx               # Root layout
│   │   └── page.tsx                 # Main page
│   ├── components/
│   │   ├── Header.tsx               # Navigation
│   │   ├── HeroSection.tsx          # Input form
│   │   ├── LoadingSection.tsx       # Progress tracking
│   │   └── ResultsSection.tsx       # Report display
│   ├── lib/
│   │   ├── api.ts                   # API client
│   │   └── types.ts                 # TypeScript types
│   ├── package.json                 # NPM dependencies
│   └── README.md                    # Frontend docs
│
├── deal_copilot/                     # Python Backend
│   ├── api/                         # FastAPI Backend
│   │   └── main.py                  # REST API server
│   ├── agents/                      # AI Research Agents
│   │   ├── deep_research_agent.py        # Gemini + Tavily
│   │   └── deep_research_agent_openai.py # OpenAI
│   ├── config/                      # Configuration
│   │   ├── config.py                # Gemini + Tavily
│   │   └── config_openai.py         # OpenAI
│   ├── cli/                         # Command-line tools
│   ├── examples/                    # Example scripts
│   ├── docs/                        # Documentation
│   └── output/                      # Generated reports
│
├── run_fullstack.sh                 # Run frontend + backend
├── run_server.sh                    # Backend only
├── requirements.txt                 # Python dependencies
├── .env                            # API keys
├── README.md                        # This file
└── GETTING_STARTED.md              # Setup guide

Documentation

Start Here

• Getting Started - Complete setup guide (5 minutes)
• Quick Reference - Commands cheat sheet

Full Documentation

• Full-Stack Guide - Web app architecture & API docs
• Project Summary - Complete project overview
• Comparison Guide - OpenAI vs Gemini + Tavily
• OpenAI Version - OpenAI-specific docs
• Gemini Version - Gemini + Tavily docs

What It Does

The Deep Research Agent produces investor-grade research reports with:

1. Market Overview - Market sizing, business models, dynamics, drivers/risks
2. Competitor Overview - Competitive landscape, positioning, moats
3. Company/Team Overview - Company analysis, team background, recent news

All with inline citations from web sources!

Usage

Run with Command Line

# OpenAI version
python -m deal_copilot.cli.main_openai \
  --company "Bizzi" \
  --sector "SaaS" \
  --region "Vietnam" \
  --website "https://bizzi.vn/en/"

# Gemini + Tavily version
python -m deal_copilot.cli.main \
  --company "Bizzi" \
  --sector "SaaS" \
  --region "Vietnam" \
  --website "https://bizzi.vn/en/"

Use Programmatically

from deal_copilot.agents.deep_research_agent_openai import DeepResearchAgentOpenAI

agent = DeepResearchAgentOpenAI()
report = agent.generate_full_report(
    company_name="Bizzi",
    website="https://bizzi.vn/en/",
    sector="SaaS",
    region="Vietnam"
)

print(agent.format_report_as_text(report))

API Keys Required

For OpenAI Version (Simpler):

• OpenAI API Key: Get at https://platform.openai.com/api-keys

For Gemini + Tavily Version (Cheaper):

• Google Gemini API Key: Get at https://makersuite.google.com/app/apikey
• Tavily API Key: Get at https://tavily.com/ (1000 searches/month free)

Add to .env file:

# OpenAI version
OPENAI_API_KEY=your_key_here

# Or Gemini + Tavily version
GOOGLE_API_KEY=your_key_here
TAVILY_API_KEY=your_key_here

Which Version to Use?

Feature	Gemini + Tavily	OpenAI
Setup	2 API keys	1 API key
Cost	~$0.10-0.30/report	~$0.50-1.00/report
Control	Full search control	No control
Free Tier	Yes	No

Recommendation: Start with OpenAI for simplicity, switch to Gemini + Tavily for production.

Development

# Install in development mode
pip install -e .

# Run tests (coming soon)
pytest

# Format code
black deal_copilot/

License

This is a proof-of-concept for demonstration purposes.

Contributing

See documentation in deal_copilot/docs/ for details.

---

Built with love using LangChain, Gemini, OpenAI, and Tavily