Podcast Buddy
An AI-powered news-to-podcast platform that transforms live news into natural two-host podcast episodes using Gemini, Edge TTS, and FFmpeg.
Podcast Buddy is an AI-powered platform that automatically converts live news into engaging podcast episodes. The system fetches trending news, summarizes articles using Gemini, generates natural conversations between two AI hosts, and produces high-quality podcast audio with minimal user interaction.
The platform automates the entire content generation pipeline — from news aggregation and summarization to dialogue generation, speech synthesis, and audio post-processing. By combining Large Language Models with speech synthesis and audio engineering tools, Podcast Buddy delivers realistic podcast episodes that sound natural and conversational.
Stack
- Frontend: Next.js, TypeScript, Tailwind CSS
- Backend: FastAPI
- AI Model: Gemini
- Speech Synthesis: Microsoft Edge TTS
- Audio Processing: FFmpeg
- APIs: REST API Architecture
- Data Source: Live News APIs
Notable bits
Podcast Buddy automates the complete podcast creation workflow. News articles are summarized using Gemini, converted into conversational scripts between two virtual hosts, synthesized into natural speech with Edge TTS, and merged into polished podcast episodes using FFmpeg.
The backend exposes REST APIs that seamlessly connect AI services with the frontend, enabling users to generate complete podcast episodes from live news in just a few steps.
Highlights
- Built an AI-powered news-to-podcast generation platform.
- Integrated Gemini for news summarization and dialogue generation.
- Used Microsoft Edge TTS for natural multi-speaker voice synthesis.
- Automated audio processing and episode generation using FFmpeg.
- Designed REST APIs connecting AI backends with responsive frontend interfaces.
- Converts live news into complete two-host podcast episodes automatically.