CompanyScout

VC-grade company research tool with AI-powered analysis, source citations, and smart caching.

Features

• Instant Company Research — Enter any company URL, get a structured investment brief
• Investment Scorecard — AI-generated scores (1-10) across 6 dimensions with reasoning
• Source Citations — Every claim links back to its source page for verification
• Smart Caching — Research is saved locally; instant reload for previously analyzed companies
• Hybrid Search Chat — Ask follow-up questions using semantic + keyword search
• Deal Memo Generator — One-click investment memo with thesis, risks, and recommendation
• Diligence Questions — Auto-generated questions specific to each company

Quick Start

1. Clone & Install

git clone https://github.com/suhaasteja/CompanyScoutAgent.git
cd CompanyScoutAgent
pip install -r requirements.txt

2. Set up API Key

Option A: Create .env file

cp .env.example .env
# Edit .env and add your OpenAI API key

Option B: Enter in UI

• Launch the app and enter your key in the Settings panel

3. Run

streamlit run app.py

Open http://localhost:8501 and try: linear.app, supabase.com, or posthog.com

How It Works

Enter URL → Discover Pages → Scrape Content → LLM Extraction → Display
                                    ↓
                            Save to SQLite + ChromaDB
                                    ↓
                            Instant reload next time

Tech Stack

Component	Technology
UI	Streamlit
LLM	OpenAI GPT-4o-mini
Embeddings	OpenAI text-embedding-3-small
Vector DB	ChromaDB
Database	SQLite
Scraping	requests + BeautifulSoup

Research Depth Options

Mode	Pages	Speed
Quick scan	5	~10s
Standard	10	~20s
Deep dive	18	~40s

Data Extracted

• Company name & description
• Industry & business model
• Founders & team size
• Tech stack
• Key products
• Funding stage & investors
• Hiring signals
• Competitive advantages
• Potential risks

All fields include source URLs for verification.

Investment Scoring

Each company is scored (1-10) on:

Dimension	What it measures
Market	Size & growth of the market
Team	Founder/team quality indicators
Product	Uniqueness & defensibility
Model	Revenue model clarity
Timing	Market timing & trends
Overall	Holistic investment attractiveness

Hover over any score to see reasoning and sources.

Limitations

• Website-only data — No financials, no growth metrics, no private data
• Cloudflare-protected sites — Some sites block scraping (e.g., openai.com, stripe.com)
• LLM judgment — Scores are AI-generated, not financial advice

Sites That Work Well

linear.app, supabase.com, posthog.com, railway.app,
vercel.com, temporal.io, modal.com, replicate.com,
huggingface.co, anthropic.com, notion.so, figma.com

Project Structure

CompanyScoutAgent/
├── app.py              # Main Streamlit application
├── company_analyzer.py # LLM-powered data extraction
├── storage.py          # SQLite + ChromaDB storage
├── requirements.txt    # Python dependencies
├── .env.example        # API key template
└── data/               # Local database (gitignored)

Contributing

PRs welcome! Areas for improvement:

• External data sources (Crunchbase, LinkedIn, GitHub)
• Company comparison mode
• Batch analysis
• Export to Notion/Airtable

License

MIT