This project provides a highly extensible, multi-model AI Telegram bot deployed on Cloudflare Workers, targeting developers and power users seeking a scalable and performant AI assistant. It supports numerous AI providers (OpenAI, Gemini, Claude, Groq, Azure) and features like conversation memory, image generation/analysis, and multilingual support, all accessible via a familiar Telegram interface.

How It Works

The bot leverages Cloudflare Workers for edge deployment, enabling low-latency responses. It uses TypeScript for development and a modular API structure to integrate various AI models. Conversation history is managed with Redis for efficient caching. Image generation utilizes DALL-E and Cloudflare's Flux, with an optional prompt optimization layer via an external API.

Quick Start & Requirements

• Install: Deploy to Cloudflare Workers using wrangler publish or the Cloudflare dashboard.
• Prerequisites: Cloudflare account, Telegram Bot Token, Upstash Redis database, API keys for AI services (OpenAI, Gemini, Claude, etc.).
• Setup: Requires configuring environment variables for API keys and tokens. Detailed deployment steps are available in the project's tutorial.

Highlighted Details

• Supports OpenAI, Google Gemini, Anthropic Claude, Groq, and Azure OpenAI APIs.
• Integrates with OpenAI-compatible APIs like One API and New API.
• Features text-to-image generation (DALL-E, Cloudflare Flux) and image analysis.
• Includes context memory for conversational continuity and user permission management (whitelist).
• Offers multilingual support for 8 languages.

Maintenance & Community

The project is maintained by snakeying. Links to community channels or roadmaps are not explicitly provided in the README.

Licensing & Compatibility

• License: MIT License.
• Compatibility: Permissive license suitable for commercial use and integration with closed-source applications.

Limitations & Caveats

The project requires specific external services (Cloudflare Workers, Upstash Redis) for operation. Users must manage their own API keys and be mindful of usage quotas. It's recommended to avoid configuring identical models across different API integrations (e.g., Gemini in both direct Gemini and OpenAI-compatible configurations) to prevent conflicts.