About the project
This is a high-tech Telegram chatbot built for financial advisory, deep market analytics, and educating users on the fundamentals of investing.
Unlike standard button-driven bots, the project is an AI agent powered by large language models (LLMs) that can hold natural conversations, understand complex requests, autonomously decide when to call external tools (web search, quote feeds, chart rendering), and deliver well-argued financial answers backed by live data.
User functionality
Users get an on-demand personal financial analyst available 24/7. Conversations happen in natural language—the bot tracks context and intent.
Key capabilities:
- Deep market analytics: Quotes from MOEX and Yahoo Finance, chart generation with indicators (SMA, RSI), and access to fundamentals (P/E, EPS).
- Information hub: Dividend calendar, issuer-specific news, and “smart” web search to refresh data on regulations and taxes.
- Education platform: Interactive financial literacy courses with progress tracking and answers sourced from the internal Knowledge Base (RAG).
- Financial calculators: Tools for profitability, mortgage, credit, and compound interest calculations.
Technologies and architecture
The solution is implemented on a Python stack using modern approaches to AI application development. Architecture is split across microservices (Bot, Worker, Server) for resilience and scalability.
Technology stack:
Implementation results
Key implementation highlights:
- Hybrid data processing: The system automatically switches between exchange APIs and web search to guarantee a response.
- Failure protection: Queues and background workers keep the bot responsive during heavy tasks (chart generation, LLM calls).
- High-quality visualization: Charts and tables adapt to the content for a messenger-friendly layout.
- Seamless integration: Instant authorization sync with the external customer database via API.