DEVELOPMENT OF A MULTIFUNCTIONAL TELEGRAM BOT AS A UNIFIED GATEWAY TO DIVERSE NEURAL NETWORKS

Pavlyuk Daniil Artemovich, Yakovlev Igor Vladimirovich

This article is dedicated to the development of a multifunctional Telegram bot that acts as a unified gateway to various neural network technologies. The primary objective of the research is to address the pressing problems of fragmentation and geographical restrictions faced by users of AI services. The project aims to create a centralized platform that integrates both official APIs from leading companies (OpenAI, Google, Anthropic) and numerous third-party providers under a single, intuitive interface, making advanced technologies accessible to a wide audience. To achieve this goal, a modular asynchronous architecture was implemented in Python, utilizing the Aiogram 3 framework for Telegram API interaction and an SQLite database for data storage. The system is designed for deployment in Docker containers, enabling it to function effectively as a proxy gateway. As a result, a bot with extensive multimodal capabilities was developed: it supports text dialogues, image generation, editing, and analysis, audio transcription, and text extraction from documents. Special attention is given to the project's key outcome: the creation of a full-featured administrative interface directly within Telegram. Unlike counterparts with static configurations, this solution allows an administrator to dynamically add and remove AI models, manage user access, and send broadcasts without editing the source code or restarting the application. This flexibility and scalability are the main competitive advantages, making the bot a powerful and easily manageable tool.

Keywords:Telegram bot, neural networks, API integration, unified gateway, multimodality, Aiogram, Python, Docker.