Mit einem Wikipedia-Podcast-Agenten auf Telegram mit Claude und ElevenLabs chatten

# Welcome to my Wikipedia Podcast Telegram Agent Workflow! ## This workflow has the following sequence: 1. Telegram Trigger - Receives incoming messages (text or voice) from users via Telegram bot 2. Text or Voice Switch - Routes the message based on input type (text message vs. voice message) 3. Voice Message Processing (if voice input): - Retrieval of voice file from Telegram - Transcription of voice message to text using OpenAI Whisper 4. Text Message Preparation (if text input) - Prepares the text message for the AI agent 5. Wikipedia Podcast Agent - Core AI agent that: - Researches the requested topic using Wikipedia tool - Creates a professional 5-minute podcast script (600-750 words) - Follows structured format: intro, main content, outro - Uses conversational, accessible, and enthusiastic tone 6. ElevenLabs Text to Speech - Converts the podcast script into natural-sounding audio using AI voice synthesis 7. Send Voice Response - Delivers the generated podcast audio back to the user via Telegram ## The following accesses are required for the workflow: - Telegram Bot API: Documentation - Anthropic API (Claude 4 Sonnet): Documentation - OpenAI API (for voice transcription): Documentation - ElevenLabs API (for text-to-speech): Documentation ## Workflow Features: - Multi-input support: Accepts both text and voice messages from Telegram users - Wikipedia integration: Automatically researches topics using Wikipedia data - Professional podcast generation: Creates structured, engaging 5-minute podcast episodes - Natural voice synthesis: Converts text to high-quality speech using AI - Seamless user experience: Users receive audio responses directly in Telegram chat ## Use Cases: - Educational content creation - Quick Wikipedia topic summaries in podcast format - Accessibility tool for consuming Wikipedia content - Personal learning assistant - Content creation for social media or educational purposes You can contact me via LinkedIn, if you have any questions: https://www.linkedin.com/in/friedemann-schuetzt