Transforming any article, topic, or idea into a fully produced podcast episode — instantly.
People want to stay informed, but reading long articles is exhausting. Turning any topic into a real podcast takes hours of scripting, recording, and editing. And when misinformation spreads faster than the truth, most people have no easy way to verify what they are sharing.
My ET transforms passive content consumption into an active podcast creation studio. Type a topic, paste an article, or search for something you care about — and get a fully scripted, voiced, and verified podcast episode in seconds.
One click turns any article or topic into a produced podcast episode.
- Choose your format: 2-person conversational, solo narration, interview-style Q&A, or short daily brief
- Gemini 1.5 Pro generates natural, format-specific scripts — not summaries read aloud
- ElevenLabs assigns distinct voices to each speaker for a genuine conversational feel
- Works on any content: news articles, blog posts, your own notes, or a topic you type in
- Search any topic and get AI-curated source material ready to convert
- Pulls from live news and reference content simultaneously
- Intent-aware layouts for topic deep-dives, comparisons, and quick fact-checks
- Upload a screenshot, article, or claim before you record
- Gemini 1.5 Pro analyzes it visually and textually
- Tavily cross-references against authoritative sources
- Returns a clear true, false, or misleading verdict with citations
- One-click script generation in 10+ languages
- GPT-4o-mini handles semantic accuracy, language-specific models handle cultural nuance
- Not literal translation — contextual adaptation that feels native to the listener
Nine tools available on every piece of content:
- Podcast Script, Solo Brief, Interview Format, Episode Arc
- Context Cards, Contrarian Angle, Multilingual Script
- Video Brief and Fact Score
To prevent AI hallucinations in scripts, every piece of content is grounded in real fetched source material via Tavily. Gemini shapes the language — it never invents the facts.
- Router Agent classifies user intent — search, script generation, or fact-check
- Content Agents fetch and synthesize from multiple sources in parallel
- Script Agent generates format-specific output with distinct prompting strategies per format
- Voice Agent directs ElevenLabs with pacing and speaker cues
- SHA-256 hashing caches generations in Neon Postgres via content-hash keys
- Repeat queries drop from 8 seconds to under 40ms
- Clone the repo
- Setup environment: Rename
webapp/.env.exampletowebapp/.envand add your API keys (Gemini, ElevenLabs, Tavily, Clerk) - Install and run:
cd webapp
pnpm install
pnpm run devBuilt for anyone with a story worth hearing.