GPT/Whisper/DALL-E Telegram Bot with easy deployment using Chalice to AWS Lambda

GPT/Whisper/DALL-E Telegram Bot using AWS features

• Lambda, DynamoDB (full auto-deployment)
• Fargate, ECS (manual Docker container deployment) Deprecated

Requirements

• Python 3.9+
• AWS CLI
• AWS credentials and region setup in the `~/.aws/credentials` file
• Chalice

Deployment to AWS Lambda

1. Create a Free Tier AWS account

2. Install and setup AWS CLI (region must be specified in ~/.aws/credentials. Ex.: region=eu-west-1)

3. Update the `.chalice/config.json` file with your Telegram Bot API, OpenAI API Token, and your telegram userID (get it from @userinfobot).

4. Run the deploy.sh script to deploy the Telegram Bot to AWS Lambda (or deploy.sh name for a custom config)

For voice processing you will need to do the next manually (it will be automated in the next releases):

1. Upload ffmpeg.zip to your S3 bucket
2. Create Lambda Layer using S3 URI
3. Attach layer to your Lambda function

Features

• Currently used openAI model: gpt-3.5, gpt-4 (Choose in config or by the bot's command)
• Image generation /image prompt or just ask without the command!
• Voice message transcript (Just send any voice message)
• Video transcript and translation! (Send any video to the bot, and it will transcribe and translate it to English)
• The context of your chat is saved until you use the command /clear
• The price in USD of the response is shown
• Multi-config support: keep multiple chalice configs to deploy multiple bots .chalice/name.config.json