gora.
EN
РусскийEnglishქართულიLietuvių
telegram bots · mini apps · georgia

Telegram Bot Developer in Georgia

I build production Telegram bots and Mini Apps for businesses in Georgia and abroad: order and notification flows, payments, AI assistants, admin tools and integrations with the systems you already run.

Discuss a Telegram botView relevant work ↓
  • 31,000+ messages handled in live Telegram projects
  • 3 Mini Apps and 3 bots in one production platform
  • 12 AI models in a single Telegram bot

What a bot or Mini App can do for a business, and where I have already done it.

  • Customer-facing bots

    Order status, notifications and customer conversations. In the delivery platform below, a customer bot receives events from the back end and messages the customer when an order moves.

  • Telegram Mini Apps

    Catalogues, checkout and dispatcher or driver tools inside Telegram. Sign-in uses Telegram's signed initData, validated server-side with HMAC-SHA-256, so users never create a password.

  • AI assistants in Telegram

    Claude, GPT, Gemini and Grok behind one bot, with memory, voice replies and per-plan limits. One of my bots routes between 12 models inside a single chat.

  • Support and auto-reply bots

    Rule-based AI replies, fallbacks between model providers and escalation to a human when confidence is low, plus AI moderation of channels and groups.

  • Payment flows

    Checkout inside bots and Mini Apps. The delivery platform accepts cash on delivery, an internal wallet and crypto (BTC, ETH, USDT, USDC, TRC-20) with blockchain confirmation tracking.

  • Admin bots and internal tools

    Owner and operator consoles in Telegram: switch features on and off, edit rules and limits, read history. Small teams often do not need a separate admin panel.

  • Notifications and alerts

    Priority-based push messages to customers, dispatchers and drivers, plus operational alerts for your own team, driven by events from your back end.

  • Integrations with your systems

    Bots that talk to CRMs, databases and internal APIs through REST and webhooks. In my architecture the API never knows about Telegram messages and bots never touch the database, which keeps both sides replaceable.

Two production systems with their numbers. Each links to the full case study.

Last-mile delivery platform — Miami

Multi-tenant delivery platform with a web storefront, an admin panel and Telegram as an operations channel (client under NDA).

What was done
Three Telegram Mini Apps (customer sales, dispatcher, driver) and three Python bots on aiogram 3, connected to a NestJS API through Redis pub/sub. Real-time order and driver-location events reach dispatchers and drivers through Socket.IO.
Problem solved
Give customers, dispatchers and drivers each a fast mobile surface inside Telegram, without coupling the back end to Telegram's message handling.
Technology
  • aiogram 3
  • Telegram WebApp HMAC
  • Vite 7 + React 19
  • NestJS
  • Redis pub/sub
  • Socket.IO
  • Docker
In numbers
  • 3 Telegram Mini Apps and 3 Python bots in one monorepo of 11 apps
  • Mini App sign-in by HMAC-SHA-256 validation of Telegram initData
  • Each bot runs in its own Docker container in production

Read the case study →

GoraBot — multi-AI assistant

A Telegram assistant with access to 12 models from four AI providers in one chat.

What was done
Model chains where four AIs process a task one after another, debates between models with a judge, project contexts with their own history, voice replies and Free, Basic and Pro tiers with limits enforced in the database.
Problem solved
Switching between four AI providers' own apps loses context and time. One chat keeps the conversation and puts every model under a single rate limit and billing logic.
Technology
  • python-telegram-bot 21
  • Anthropic, OpenAI, Google, xAI SDKs
  • SQLite
  • FastAPI
  • systemd
In numbers
  • 12 models in one chat: Claude, GPT, Gemini and Grok
  • 56 users and 1,783 generations in the production database on the publication date
  • 14 active project contexts, each with its own conversation history

Read the case study →

Bots
Python aiogram 3 · python-telegram-bot · TypeScript telegraf
Mini Apps
Telegram WebApp · HMAC initData validation · Vite + React · Tailwind CSS
Back end
NestJS · FastAPI · PostgreSQL · Redis pub/sub · Socket.IO
AI
Anthropic Claude · OpenAI GPT · Google Gemini · xAI Grok
Other channels
WhatsApp Business API · Twilio A2P SMS
Operations
Docker Compose · systemd · nginx · GitHub Actions

From first message to a bot running in production, with one person accountable.

  1. 01 · First message

    Write to me on Telegram. In your own words, even if you don't know the technical terms. I care more about understanding what you want to build than what you call it.

  2. 02 · Reading between the lines

    Within a day or two I understand the task and give you an estimate of timeline and budget. If I see it isn't my case — I say so right away, no hedging.

  3. 03 · Lock-in

    Price and timeline — in writing. 50% upfront. After that things change only by mutual agreement, and rarely.

  4. 04 · Work

    I show interim results at my own pace. At the end — handover with documentation and access. First month of support is included by default.

  • Direct communication with the developer. No agency middle layer and no project manager relaying messages.
  • Architecture, bot logic, Mini App front end, back end and deployment from one person.
  • Production experience: 31,000+ messages handled across live Telegram projects.
  • I work with existing systems, not only greenfield projects: bots connect to your current API or database.
  • Source code and credentials are yours from day one. If you later want another developer, they can pick it up.
  • Based in Tbilisi, working internationally.

I am based in Tbilisi and work with Georgian businesses as well as clients abroad. I communicate in English and Russian, with basic Georgian.

For Georgian businesses I can integrate local payment and operational systems where required, including workflows involving TBC Bank or Bank of Georgia. The agreement with the bank and its verification process stay on your side; I build the integration against the bank's API and agree its scope before I quote.

There is no standard price for a bot, because a notification bot and a Mini App with checkout and an admin panel are very different projects. The cost depends on whether you need a bot only or also a Mini App, which integrations are involved (payments, CRM, your own API), whether there is an admin interface, and whether AI features are included.

AI features add running costs that go to the model provider, not to me. For a sense of scale: in GoraBot's production database, 1,783 generations across 12 models used 5.4M input and 1.06M output tokens and cost $43.3 in API fees.

Send a short description of the scenario on Telegram or by email. I reply with a written estimate of timeline and budget, usually within a day or two, and fix both in writing with 50% upfront.

How much does a Telegram bot cost?
It depends on scope: a bot only or a bot plus Mini App, integrations, an admin interface and AI features. I do not publish a price list; I send a written estimate of timeline and budget within a day or two of your first message. AI usage is billed by the model provider separately.
Can you build a Telegram Mini App?
Yes. I have built three Mini Apps for one production platform: a customer catalogue with checkout, a dispatcher console and a driver app. Users sign in through Telegram itself, validated on the server with HMAC-SHA-256, so no passwords are needed.
Can a Telegram bot accept payments?
Yes, through a payment provider or a hosted checkout, depending on your market. In one production platform, checkout supports cash on delivery, an internal wallet and cryptocurrency (BTC, ETH, USDT, USDC, TRC-20). For Georgian businesses I can integrate flows involving TBC Bank or Bank of Georgia where their APIs are available.
Can you integrate an existing CRM or database?
Yes. Bots connect to existing systems through REST APIs and webhooks. I keep the bot layer separate from your data layer, so changing either one later does not break the other.
Can you add AI responses to a bot?
Yes. I have built bots on Claude, GPT, Gemini and Grok, including fallbacks between providers so a single outage does not stop replies, and escalation to a human when confidence is low. You choose the model and the provider account.
Can you maintain or extend a bot someone else built?
Yes, after a short review of the code and hosting. If the existing bot is not worth keeping, I will tell you in the first reply rather than rewrite it by default.
Do you work with Georgian companies?
Yes. I am based in Tbilisi and work with clients in Georgia and internationally, in English and Russian, with basic Georgian. I can also integrate local payment and operational systems where required.
Where does the bot run and who owns it?
You own the bot token, the code repository and the data. I usually deploy to a VPS with Docker or systemd behind nginx, or to your own infrastructure, and hand over documentation and access at the end. The first month of support is included.

Describe the task in your own words. I respond within 24 hours, usually the same day on weekdays.

Elsewhere: GitHub·Telegram channel·Reddit·X