📷 Demo

teacher_bot

AI-powered Telegram bot that helps students solve educational tasks via text and photos.

---

🚀 Features

• Solve tasks with plain text
• Solve tasks from photos (vision + reasoning)
• Daily token limit and notification
• Minimal vision token usage for cost-efficiency

---

🧠 Architecture

Text flow
Text → Mistral → Telegram response

Photo flow
Photo → Gemini (vision) → Extracted text → Mistral → Telegram response

Gemini is used only for vision (OCR).
Mistral handles reasoning and answers.

---

🛠 Tech Stack

• Python 3.10+
• python-telegram-bot
• google-generativeai (Gemini)
• mistralai (Mistral)
• python-dotenv
• SQLite (built-in)

---

📦 Installation (local)

1) Create and activate virtual environment

macOS / Linux

python3 -m venv venv
source venv/bin/activate

Windows

python -m venv venv
venv\Scripts\activate

2) Install dependencies

pip install -r requirements.txt

If requirements.txt is missing, install manually:

pip install python-telegram-bot google-generativeai mistralai python-dotenv

---

🔐 Configuration (.env)

Create a .env file in the project root:

BOT_TOKEN=your_telegram_bot_token
MISTRAL_API_KEY=your_mistral_api_key
GEMINI_API_KEY=your_gemini_api_key
ADMIN_IDS=123456789,987654321

Notes:

• BOT_TOKEN is required
• MISTRAL_API_KEY is required for solving tasks
• GEMINI_API_KEY is required for photo solving
• ADMIN_IDS is optional (comma-separated Telegram user IDs)

▶ Run the bot

python main.py

---

📁 Project Structure

• main.py — application entry point
• app/config.py — environment configuration
• app/handlers.py — Telegram message handlers
• app/services.py — Mistral integration and business logic
• app/vision.py — Gemini photo text extraction
• app/limits.py — daily token limit logic
• app/db.py — SQLite database logic

---