This n8n workflow creates an intelligent voice and text chatbot integrated with Telegram, enabling seamless voice recognition, AI-powered response generation, and voice synthesis. The process begins when a user sends either a voice message or a text in Telegram. If a voice message is received, the workflow retrieves the voice file from Telegram and transcribes it into text using ElevenLabs speech-to-text service. The transcribed text, alongside any textual messages received, is then processed through AI language models like Google Gemini or Groq to generate a context-aware response.

The AI’s reply is then converted back into speech using ElevenLabs Text-to-Speech (TTS) for voice messages or sent as plain text for text responses. The workflow also manages session memory for continuous conversations and supports language detection for multilingual interactions. Practical use cases include virtual assistants, customer support bots, or language learning tools, providing users with a natural and engaging conversational experience whether through voice or text.

Node Count	11 – 20 Nodes
Nodes Used	@elevenlabs/n8n-nodes-elevenlabs.elevenLabs, @n8n/n8n-nodes-langchain.agent, @n8n/n8n-nodes-langchain.lmChatGoogleGemini, @n8n/n8n-nodes-langchain.lmChatGroq, @n8n/n8n-nodes-langchain.memoryBufferWindow, aggregate, httpRequest, if, set, stickyNote, switch, telegram, telegramTrigger