Telegram · May 2026

Telegram bot for construction control — AI, geo, media

An AI assistant in Telegram for managing technical supervision on construction sites. The bot runs scheduled morning and evening reports, walks users through checklists with verified geolocation, archives photos, videos and documents to an external server with a clear folder hierarchy, and helps analyze photo reports via GPT-4o Vision.

What was built:

Reporting pipeline: scheduled morning and evening reports (APScheduler, UTC+4 timezone, Asia/Dubai)
Checklists with geolocation verification: photos, videos, comments; GPS extraction from image EXIF (Pillow, pillow-heif)
Archiving of media and documents to an external server with an 'Object → Date → Report type' hierarchy (async via aiohttp/aiofiles)
AI analysis of photo reports using OpenAI GPT-4o Vision
Consolidated checklist registry with search and export by period (pandas, openpyxl)
Onboarding of new participants: a welcome pack with rules and work schedule
Three interface languages (EN/UK/RU), quality control of construction processes

Stack: Python 3.10+, python-telegram-bot 20.7, OpenAI GPT-4o Vision, SQLite3, pandas/openpyxl/xlrd, APScheduler, aiohttp/aiofiles, Pillow/pillow-heif, pytz; deployment Docker + Railway/VPS.

Role: design and development of the construction-control Telegram bot with AI and geodata (backend, Python).

Project screens

Screen 1
Screen 2
Screen 3
Screen 4