AI Customer Support Telegram Bot — Full Source Code

A complete, production-ready customer support bot for Telegram, combining an AI assistant, a lightweight knowledge base, and human operator hand-off into one clean codebase. Built for agencies and freelancers who ship support bots for clients, or for anyone who wants to automate their own Telegram support channel without building the infrastructure from scratch.

How it works:
Telegram → AI → Knowledge Base → PostgreSQL → Admin Panel (in-chat)

Key features:

• 🧠 AI-powered replies through any OpenAI-compatible API (OpenAI, OpenRouter, Groq, or a self-hosted model) — swap providers by changing one URL

• 📚 Built-in FAQ / knowledge base with keyword search that grounds AI answers in your actual policies, no vector DB required

• 🙋 One-command operator hand-off — a customer types /operator, the conversation drops into your support team's Telegram group, and admins reply with a native Telegram "Reply" — no extra tooling

• 🗂 Full conversation history — every message (user, AI, and operator) logged to PostgreSQL

• 🛠 Admin commands, right in Telegram — statistics, FAQ management, broadcast messaging, per-user history, and more

• ⚙️ Live-editable system prompt — change the AI's tone, rules, or business context with /setprompt, no redeploy needed

• 📊 Built-in statistics — users, messages, new signups, open operator requests

• 🐳 Docker & docker-compose — the entire stack (bot + PostgreSQL) starts with a single command

• 🔐 .env.example with every variable documented, plus a detailed README and step-by-step setup guide

Tech stack: Python 3.12, aiogram 3 (async), SQLAlchemy 2.0 + asyncpg, PostgreSQL 16, Docker.

What's included:

• Full, commented source code (33 files, clean modular structure: handlers, database models, AI layer, utilities)

• Dockerfile + docker-compose.yml for instant deployment

• .env.example with all configuration variables

• Sample FAQ dataset + seed script

• Detailed README and a full PDF setup/usage guide

Great starting point for: e-commerce support bots, SaaS help desks, internal IT support bots, or as a white-label base you customize and resell to your own clients.