Telegram GPT Worker: a multi-model AI Telegram bot deployed on Cloudflare Workers

General Introduction

GPT-Telegram-Worker is a multi-model AI Telegram robot based on Cloudflare Workers, supporting multiple APIs such as OpenAI, Claude, Azure, etc. The project is developed in TypeScript, with a modularized design for easy expansion to provide a fast and scalable service experience. The robot is equipped with intelligent dialog, image generation, image analysis and other functions, and supports multiple languages to meet diversified needs.

Function List

• Multi-model support: Integrate multiple AI models such as OpenAI, Google Gemini, Anthropic Claude, Groq and Azure OpenAI.
• Intelligent dialog: with contextual memory capability to ensure smooth and natural dialog.
• Image Generation: Supports textual descriptions to generate images, using DALL-E and Cloudflare. Flux Technology.
• Image analysis: support users to upload images and analyze them intelligently, using OpenAI or Google Gemini Model.
• Multi-language support: Built-in i18n function supports 8 languages.
• User rights management: control access rights through the whitelist function to enhance security.
• High-performance deployment: Leverage the edge computing power of Cloudflare Workers for rapid response.
• Efficient data management: Redis is used for data caching and management to ensure efficient processing.
• Flux Cue Optimization: Optimize image-generated cues for Flux models via an external API.

Using Help

Installation process

1. preliminary::
   • Cloudflare Account
   • Telegram Account and Bot Token
   • Upstash Redis database (requires Eviction to be enabled)
   • API key for at least one AI service
2. Getting Started::
   • Cloning Project Warehouse:git clone https://github.com/snakeying/GPT-Telegram-Worker.git
   • Configure the necessary environment variables: in the .env file to set up information about Cloudflare, Telegram, and Redis.
   • Deploy to Cloudflare Workers:

     npm install -g @cloudflare/wrangler
     wrangler login
     wrangler init telegram-bot
     cp dist/index.js telegram-bot/
     wrangler publish
     

   • Set up a Telegram Webhook:

     curl -F "url=https://your-worker.your-subdomain.workers.dev" https://api.telegram.org/bot<YOUR_BOT_TOKEN>/setWebhook
     

Instructions for use

1. Activate the robot.::
   • utilization /start Command to start the robot.
   • utilization /language Switch language.
   • utilization /switchmodel Toggle the AI model.
   • utilization /new Start a new conversation.
   • utilization /history Get a summary of the conversation history.
   • utilization /help Get help information.
2. Image Generation::
   • utilization /img command to generate an image (DALL-E).
   • utilization /flux command to generate an image (Cloudflare Flux).
3. image analysis::
   • After uploading the image, the robot automatically performs an intelligent analysis and returns the results.
4. Rights Management::
   • Ensure security by controlling user access rights through the whitelisting feature.
5. data management::
   • Use Redis for data caching and management to ensure efficient processing.

Detailed Operation Procedure

1. Configuring Environment Variables::
   • In the project root directory, create the .env file, add the following:

     TELEGRAM_BOT_TOKEN=<Your Telegram Bot Token>
     CLOUDFLARE_ACCOUNT_ID=<Your Cloudflare Account ID>
     REDIS_URL=<Your Redis URL>
     

   • Make sure that all variables are properly configured.
2. Deploying to Cloudflare Workers::
   • Use the Wrangler CLI to deploy the project:

     wrangler publish
     

   • Configure environment variables in the Cloudflare Dashboard.
3. Setting up the Telegram Webhook::
   • Setting up a Webhook using the Telegram Bot API:

     curl -F "url=https://your-worker.your-subdomain.workers.dev" https://api.telegram.org/bot<YOUR_BOT_TOKEN>/setWebhook
     

4. Use of robots::
   • After starting the robot, you can use various commands to interact with the robot, generate images, analyze pictures, switch languages and models, and so on.