Build a fast, scalable, and production-grade Telegram Bot in 5 minutes using AWS Lambda, FastAPI, and the Serverless Framework — with full support for local development via uvicorn and ngrok.

• For pet projects, side projects, and production bots

• Manual server hosting
• Complicated scaling
• Maintenance headaches

This project solves that.

✅ SnapStart dramatically reduces Lambda cold start times..

✅ Horizontal scaling is automatic by AWS Lambda.

✅ Local development is effortless with uvicorn and ngrok.

✅ From a small side-project to millions of users — this architecture just works

✅ Parsed requests are fully typed with Pydantic v2, making your code cleaner and safer, just add business logic.

⚡ Features

• Blazing-fast response time (with AWS Lambda SnapStart)
• Fully serverless: no servers to manage
• Easily deployable in under 5 minutes
• Supports local FastAPI debug with uvicorn and ngrok
• Scales horizontally by design
• Fully typed with Pydantic v2
• Plug-and-play with Telegram Webhook API

🛠️ Technical Stack

• AWS Lambda – serverless compute
• API Gateway – to expose FastAPI as webhook endpoint
• FastAPI – async Python web framework
• Mangum – ASGI adapter for AWS Lambda
• httpx – async HTTP client to call Telegram API
• Serverless Framework – deploy & manage the stack
• Loguru – structured logging
• Pydantic v2 – data parsing & validation
• Docker - for collecting dependencies

src/ ├── app/ │ └── dispatcher.py # Business logic ├── models/ │ └── telegram_request.py # Typed Telegram webhook model ├── infrastructure/ │ └── telegram/ │ └── telegram_api.py # Wrapper around Telegram Bot API ├── entry_point_fastapi_app.py # FastAPI + Mangum Lambda entrypoint ├── settings.py # ENV-based config 

There is a single model for all Telegram requests, which is typed with Pydantic v2. It includes fields that Telegram can send, so you can easily access any data in your bot logic.

# part of src/models/telegram_request.py class Message(BaseModel): message_id: int from_: User = Field(..., alias="from") chat: Chat date: int text: Optional[str] = None entities: Optional[List[MessageEntity]] = None media_group_id: Optional[str] = None photo: Optional[List[Photo]] = None document: Optional[Document] = None class MyChatMember(BaseModel): chat: Chat from_: User = Field(..., alias="from") date: int old_chat_member: ChatMemberStatus new_chat_member: ChatMemberStatus class TelegramRequest(BaseModel): update_id: int message: Optional[Message] = None my_chat_member: Optional[MyChatMember] = None class Config: populate_by_name = True

1. Open Telegram
2. Search for @BotFather

3. Send the command start and /newbot and follow the instructions
4. Once your bot is created, BotFather will give you a token.

5. Save this token to your local environment so the bot can use it.

export TELEGRAM_TOKEN=8197297946:AAHVUKxdVq1b15O9-JJ2Xp6SyWVopbLdL5s

https://docs.aws.amazon.com/cli/v1/userguide/cli-configure-files.html

git clone https://github.com/eslazarev/aws-lambda-telegram-bot.git cd aws-lambda-telegram-bot

Edit the dispatcher.py file to implement your bot's business logic. This is where you handle incoming messages and commands.

from src.infrastructure.telegram.telegram_api import send_message from src.models.telegram_request import TelegramRequest from loguru import logger async def telegram_dispatcher(telegram_request: TelegramRequest) -> None: """  Dispatcher for handling Telegram requests.  :param telegram_request: TelegramRequest  :return: None  """ if telegram_request.message.text == "/start": logger.info(f"Received /start command from user {telegram_request.message.from_.id}") await send_message( chat_id=telegram_request.message.chat.id, text="Welcome to the bot! How can I assist you today?", reply_to_message_id=telegram_request.message.message_id, ) return # Handle other commands or messages if telegram_request.message.text: logger.info(f"Received message from user {telegram_request.message.from_.id}: {telegram_request.message.text}") await send_message( chat_id=telegram_request.message.chat.id, text=f"I received your message.\nBut I don't know how to respond yet.", reply_to_message_id=telegram_request.message.message_id, ) return

npm install -g serverless

npm install

pip install -r requirements-local.txt

(https://ngrok.com/downloads)[https://ngrok.com/downloads]

uvicorn src/entry_point_fastapi_app:app --reload

ngrok http 8000

Use the ngrok URL to set the webhook for your bot. Replace <ngrok_url> with the actual URL provided by ngrok.

curl -X POST "https://api.telegram.org/bot$TELEGRAM_TOKEN/setWebhook" -d "url=https://<ngrok_url>/webhook"

Send a message to your bot in Telegram. You should see the request being logged in your terminal where you run the FastAPI server.

serverless deploy -r eu-central-1 --param=$TELEGRAM_TOKEN

Use the API Gateway URL provided by Serverless Framework to set the webhook for your bot. Replace <api_gateway_url> with the actual URL.

curl -X POST "https://api.telegram.org/bot$TELEGRAM_TOKEN/setWebhook" -d "url=https://<api_gateway_url>/"

Send a message to your bot in Telegram. You should see the request being logged in your AWS CloudWatch logs.