HUMANIKRONOver the past while I have built dozens of Telegram bots for small business — cafés, shops, salons. Here is what my process looks like and why it is fast.
1. A spec in 10 minutes
I do not write reams of documents. I ask the client three things: what the bot should do, who will use it and where to send the orders. That is enough to quote a fixed price and deadline — no surprises along the way.
2. A stack that does not let you down
- Python + aiogram — fast and stable for bots of any size.
- FastAPI — when you need webhooks, an admin panel or integrations.
- Groq / LLaMA — for AI answers to customer questions, when you need "smart" replies without an expensive API.
- SQLite — for orders and statistics; nothing extra.
3. What is always inside
A menu with categories, order taking, instant notifications to the owner in Telegram, hours and contacts. In higher tiers — AI replies, search, statistics, an admin panel and payments (Telegram Stars or Mono).
4. Deployment and support
For testing the bot runs locally, for continuous operation — a small VPS (~$3–5/mo). 7 days of support are always included.
Why a fixed price
Because I already know how long each block takes. The client pays for the result, not for my experiments. If you need a bot — message me on Telegram, and in a couple of days it will be running.