Addis Laptop

Addis Laptop is a Python-based Telegram bot and web app that aggregates laptop listings from Ethiopian Telegram tech channels, extracts specs using LLMs, and provides intelligent search and recommendations.

Features

• Browse latest laptop listings (pagination)
• Filter by brand, price, RAM, screen size
• AI-powered recommendations by use case and budget
• Natural language search (English & አማርኛ)
  • "Dell laptop under 100k"
  • "Leonovo laptop 16GB RAM"
  • "ለቢሮ ሥራ የሚሆኑ ላፕቶፖች፣ ከ50ሺ በታች"
• Voice search (English & አማርኛ)

Architecture

View architecture diagram

Live Demo

• Telegram Bot: @AddisLaptopBot
• Web App: https://laptops.balewgize.app/

Video

video_1.mp4

Quick Start

Prerequisites

• Python 3.11+
• Telegram API credentials (my.telegram.org)
• Telegram Bot token (@BotFather)
• OpenRouter API key (openrouter.ai)
• Docker

Installation

git clone https://github.com/balewgize/addis-laptop.git
cd addis-laptop
python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env
# Edit .env with your credentials

Configuration

# .env
TELEGRAM_API_ID=your_api_id
TELEGRAM_API_HASH=your_api_hash
TELEGRAM_BOT_TOKEN=your_bot_token
OPENROUTER_API_KEY=your_openrouter_key

# Optional (for voice search)
ELEVENLABS_API_KEY=your_elevenlabs_key

Data Extraction

# Option 1: From Telegram exports (recommended)
# Scrape manually from exported JSON (to avoid account ban)
# Export messages from Telegram Desktop (JSON format)
# Place the JSON file in the data/exports directory
# Filename must be channel username without @ symbol
# e.g. data/exports/channel_username.json
python -m scripts.sync_json

# Sync specific channel
python -m scripts.sync_json data/exports/channel_username.json

# Option 2: Using Telethong Sync channels (scrape laptop listings)
# Note: It will require you to login to your Telegram account
# WARNING: This may get your account banned if used frequently
python -m scripts.sync_channels

Run with Docker (local)

docker compose -f docker-compose-local.yml up --build

Run with Docker (prod)

docker compose up -d --build

Run each service alone (optional)

# Start the bot
# python -m bot.core

# # Run Streamlit user app
# streamlit run web/user.py

# # Run Streamlit admin app
# streamlit run web/admin.py

Accessing locally

• Telegram Bot: @your_bot_user_name
• Web User: http://localhost:8501/
• Web Admin: http://localhost:8502/

Project Structure

Project Structure
addis-laptop/
├── bot/          # Telegram bot (handlers, parsing, voice)
├── core/         # Business logic (scraping, extraction, recommendations)
├── web/          # Streamlit apps (user + admin)
├── scripts/      # CLI utilities for syncing data
├── data/         # SQLite database
└── tests/        # Test suite

Tech Stack

• Language: Python 3.11+
• Bot: python-telegram-bot
• Web: Streamlit
• Scraping: Telethon (Telegram API)
• Database: SQLite + SQLModel
• LLM: OpenRouter API (Claude/Gemini)
• Deployment: Docker + Caddy

License

MIT