Voice AI Assistant: Telegram + Deepgram + OpenAI + Supabase

This n8n workflow creates a fully functional AI audio assistant integrated with Telegram, enabling natural voice interactions. It receives voice messages from users via a Telegram bot, downloads the audio files, and transcribes them accurately using Deepgram's speech-to-text (STT) technology. The transcribed text is then processed by an OpenAI GPT agent that accesses a Supabase knowledge base for informed responses, maintaining conversation context through memory management for coherent, multi-turn dialogues.

The workflow generates intelligent, context-aware responses and converts them back to natural-sounding speech using Deepgram's text-to-speech (TTS). The resulting audio file is sent directly back to the user in Telegram, providing a seamless hands-free experience. This eliminates typing barriers, supports multilingual interactions, and ensures 24/7 availability.

Ideal for businesses in customer service, content creation, accessibility services, or any scenario needing voice-based AI support. Benefits include reduced response times, enhanced user engagement, scalability without human intervention, and cost savings on support staff. Perfect for global teams breaking language barriers or hands-free operations in dynamic environments.