Telegram Mini App for learning English, AI dictionary

Full-stack portfolio project: personal AI dictionary of the English language in Telegram.

The user sends a word or phrase to the bot (via text or voice) or adds them in the Mini App — the system generates a learning card: translation, meaning, examples, synonyms, CEFR level. Words are stored in a personal database, can be searched, filtered by status (new → learning → learned), and reviewed.

What has been implemented:
• Telegram bot (aiogram 3): text, voice (Whisper), commands /start, /help, /app
• Telegram Mini App (React + TypeScript + Vite): mobile-first UI, dark/light theme, animations
• Backend on FastAPI: layered architecture (routes → services → repositories)
• PostgreSQL + Alembic migrations
• OpenAI integration for structured cards and follow-up questions
• Daily streak, goals, and activity calendar
• Linking words to movies/series (TMDB) and music (MusicBrainz)
• Security: validation of Telegram initData (HMAC), no trust in client-side user ID in production
• Docker Compose for local development, deployment on Railway, CI on GitHub Actions

Stack: Python 3.11, FastAPI, aiogram, SQLAlchemy 2 async, PostgreSQL, React 18, TypeScript, OpenAI API, Docker, Railway.

The project demonstrates the full cycle: from idea and architecture to deployment, testing, and production-ready settings.

#telegram #telegrambot #miniapp #python #fastapi #react #typescript #openai #postgresql #fullstack #webdevelopment #ai #english #vocabulary