I want a fully-featured Telegram bot written in Python that can handle customer support chats through the GPT-4o-mini model and also run face-recognition analysis on images by calling GPT-4.1. Key workflow • Users join a mandatory channel before any commands work. • Free users get a small daily question quota; Premium users enjoy unlimited requests. • A referral counter automatically unlocks 14 days of Premium after 25 invited friends and 50 days after 50 friends. • All conversations, referrals, premium flags and tickets live in a single SQLite database (bot.db) for easy backup. I also need: • A lightweight support-ticket flow that lets users raise an issue and lets me reply from within the bot. • An admin panel (simple inline-menu is fine) where I can search users, grant or revoke Premium, broadcast messages, and view basic stats. • Fast, accurate face-recognition feedback whenever a user sends a photo. • Clear, well-commented code plus a step-by-step deployment guide so I can swap in my OpenAI and Telegram tokens and be running in minutes. Deliverables 1. Complete Python source (prefer aiogram but open to alternatives) 2. bot.db schema and any helper scripts for migrations or seeding test data 3. README / setup guide that works on Ubuntu VPS 4. Short screencast or screenshots proving the main flows function end-to-end My target outlay is roughly 500 000 UZS for the whole solution, payable on successful hand-over of the above items. Let me know your timetable and any relevant demos you have built previously—if everything looks good, just say “ORDER” and we can get started.