SmartExpenseBot — Telegram Personal Finance Assistant

Personal finance tracking tools often suffer from low user engagement due to complex interfaces and manual workflows. Users want a frictionless way to log expenses quickly — especially on mobile — and many prefer speaking or using their native language rather than navigating traditional apps.

Additionally, most lightweight tools lack intelligent categorization, multilingual support, and integrated income tracking, limiting their usefulness for real-world financial awareness.

SmartExpenseBot was designed as a conversational, AI-first finance assistant inside Telegram.

Using multiple specialized AI instances powered by DeepSeek, the bot understands natural language messages and voice inputs to automatically extract financial data such as amount, currency, and category.

The system provides:
‣Seamless expense and income logging through chat or voice
‣Automated categorization and currency normalization
‣Natural-language financial queries and reports
‣Smart reminders and timezone-aware scheduling
‣Multilingual UX (English, Russian, Uzbek)

The architecture ensures scalability with PostgreSQL support while maintaining privacy through local database storage options.

‣Delivered a voice-first financial tracking experience reducing friction in daily logging
‣Implemented AI-driven expense extraction and categorization improving data accuracy
‣Enabled multilingual financial assistant capabilities across three languages
‣Built a scalable architecture supporting both local and production database environments
‣Demonstrated practical integration of LLM + speech recognition in a real consumer tool