Work in Progress — Shipping Soon

Your podcast's
AI writers' room.

4 AI personas watch your live show and react in real-time. A fact-checker. A sound effects guy. A comedy writer. A cynical troll. Inspired by the Howard Stern Show. Streaming alongside the conversation as it happens.

The challenge, straight from @jason & @lonharris on This Week in Startups

The Cast

4 personas. Inspired by the Howard Stern Show.

Each persona mirrors a member of the Stern Show staff — the fact-checker, the sound guy, the comedy writer, the troll. Each runs on a different AI model. No single point of failure.

🎯

The Fact-Checker

Gary Dell'Abate · Claude Haiku + Brave Search

Monitors the conversation for factual claims and provides corrections or background data in real-time. Searches the web mid-show to verify statistics, dates, and attributions.

[FACT CHECK] "Jason just said Uber was founded in 2007. It was 2009. Again."
🔥

The Cynical Troll

Groq · Llama 70B · 120ms

The "chaotic or negative cynical" commentator. Says what the audience is thinking but won't type. Dunks with internet-brain energy.

"Oh cool, another AI wrapper. Very 2024."
🎧

Sound Effects Guy

Fred Norris · Groq · Llama 8B

Supplies background context and sound effects. Communicates through precisely timed sound cues and the occasional razor-sharp one-liner.

[record scratch] Fun fact: that company went bankrupt in 2023. [sad trombone]
😂

The Comedy Writer

Jackie Martling · Claude Haiku

Generates one-liners and jokes related to the current discussion. Setup-punchline structure, callback humor, the one who makes you spit out your coffee.

"Jason's investment thesis: if it has 'AI' in the name and the founder has a pulse, it's a yes."
Architecture

Paste a URL. Watch it together.

The video plays right inside the app. Pause it and the AI pauses too. Resume and everyone picks back up.

Watch Together

Embedded YouTube player on the left, 4 AI personas on the right. One screen, full experience. Synced play/pause controls.

Real-Time Transcription

Audio streams to Deepgram's Nova-3 model over WebSocket. Sub-300ms latency. Live transcript scrolls beneath the video.

Persona Engine

Every 90 seconds of new dialogue, all 4 personas fire in parallel. Each gets rolling context + their own memory for continuity.

Streaming Sidebar

Responses stream token-by-token via SSE into a 2×2 persona grid. Dark theme. Auto-scroll. Typing indicators.

No Platform Trap

Multi-provider by design.

Two LLM providers. Swap any model. No single dependency. Open source, MIT licensed.

Next.js 15
App Router + SSE
Deepgram Nova-3
Real-time transcription
Claude Haiku
Fact-Checker + Comedy Writer
Groq + Llama
Troll + Sound Effects Guy
Brave Search
Live fact-checking
Tailwind + shadcn
UI components

Built for a bounty.

Jason Calacanis and Lon Harris put out a challenge on This Week in Startups: $5,000 + a guest spot to whoever builds a live AI sidebar with 4 personas watching the pod in real-time. Open source. Ship it.

So we're shipping it.

Built by Seth Rininger  for @TWiStartups

🚧 Currently Building

Star the repo to get notified when we ship. It won't be long.