🤖 Local Lead Qualifier Bot

Powered by Gemma 4 (local) — Zero API costs. Fully private.

What it does

Qualifies sales leads using Google's Gemma 4 model running locally on your machine. No cloud API. No data leaves your laptop.

Send a lead's details to Telegram → get a qualification score + personalized outreach message instantly.

Demo

1. Open Telegram → search your bot
2. Send /start
3. Fill in lead details step by step
4. Get score + outreach message in seconds

Tech Stack

• Google Gemma 4 E2B (local inference)
• Ollama (model runner)
• Python
• python-telegram-bot
• Runs 100% offline after setup

Setup

1. Install Ollama

Download from https://ollama.com and install.

2. Pull Gemma 4

ollama pull gemma4:e2b

3. Clone the repo

git clone https://github.com/25Devmaker/local-lead-qualifier
cd local-lead-qualifier

4. Install dependencies

pip install python-telegram-bot ollama

5. Add your Telegram bot token

Open app.py and replace:

TOKEN = "YOUR_TELEGRAM_BOT_TOKEN"

Create a bot via @BotFather on Telegram to get your token.

6. Run

python app.py

Project Structure

local-lead-qualifier/
├── app.py              # Telegram bot entry point
├── qualifier.py        # Gemma 4 qualification logic
├── loader.py           # CSV lead loader
├── exporter.py         # Excel export
├── leads.csv           # Sample leads (bring your own)
└── requirements.txt    # Dependencies

How it works

1. User sends /start → bot asks for lead details
2. Each field is captured step by step
3. Gemma 4 analyzes the lead locally
4. Returns score (1-10), reason, outreach message
5. Results displayed in Telegram + saved to Excel

Key Features

• 100% local — no cloud APIs
• Zero API costs
• Fully private — data never leaves your laptop
• Fast — Gemma 4 runs on consumer hardware
• Simple Telegram interface
• Clean Excel output

Use Cases

• Qualifying leads from LinkedIn scraping
• Internal sales qualification
• Automated lead scoring
• Quick outreach message generation

Customization

• Change Telegram token in app.py
• Modify scoring logic in qualifier.py
• Add more lead fields if needed
• Change Excel output path in exporter.py

Troubleshooting

• Ensure Ollama is running: ollama list
• Check model is pulled: ollama pull gemma4:e2b
• Verify token in app.py
• Check Python dependencies installed

Requirements

• 8GB RAM minimum
• 2GB free disk space
• Python 3.10+
• Ollama installed

Built by

Hari Kishan Reddy H G (25Devmaker)

• Portfolio: https://25devmaker.vercel.app/
• LinkedIn: https://www.linkedin.com/in/hari-kishan-devmaker
• Fiverr: https://www.fiverr.com/s/BRYYlwd

Part of my AI projects series

This is Project 08 in my public AI build series. Check my LinkedIn for the full journey.

License

MIT

Author

25Devmaker