Caloroga's Linnet Turns One Prompt Into Editable Multitrack Shows

Caloroga Media opened its AI production studio Linnet to the public in beta on November 2, promising a full rough cut of a podcast or video from a single prompt — priced from $16 a month for a bundle the company says would cost $300 if bought separately.

Caloroga has produced roughly 60 podcasts with just two founders and a roster of AI voices. Now it's selling the workflow behind that output. Co-founder Mark Francis, a former producer at WNEW-FM in New York and later Senior Director of Talk Production at SiriusXM, frames the tool as a standards play.

"We're trying to take a really high ground on the use of AI, its content, its creation, and set a standard for doing it right," Francis said. He built Linnet with co-founder John McDermott, a former SiriusXM and Cadence13 executive. The pair started Caloroga in New York in 2023, and Linnet grew out of the network's internal workflow.

What does one prompt actually produce?

Linnet's AI layer, called "God Mode," assembles an episode from external services. In Francis's example, a prompt for a two-minute dog-walking video sends the request to Claude for a script, pulls a music bed from Jamendo, a voice from ElevenLabs, stock footage from Pexels and podcast artwork from Ideogram, then lays everything on a multitrack timeline.

That final step is the pitch. "This puts it out on a timeline where you can now edit it, make it better, swap in your own footage, swap in your voiceover if you want," Francis said. With an API key from Podigee, Transistor or Buzzsprout, God Mode can publish a show within 10 minutes, according to him. That claim is the company's own, untested independently.

The workstation carries 24 audio tracks plus video and graphics tracks, with fades, audio separation and volume automation — modeled on software like Pro Tools that producers already use. Francis dismisses the first wave of AI voice tools: "They were built by coders, not by broadcasters."

He expected a couple of months of vibe coding on Replit. It took a year and a couple thousand hours — though he counts that as a million dollars of investor money saved after nobody funded the idea.

Where does Linnet draw its lines?

Not on content limits. Francis concedes someone could use God Mode to make slop; his only guardrails are legal. An AI voice should never deliver a personal testimonial, a rule he carries over from FCC broadcasting standards. Linnet pays for the rights to assets it uses, watermarks output and follows newly introduced EU rules. "It took a lot of time and lawyers to get it right, but it sets us apart from those companies that are popping up like a dime a dozen," he said.

The company doesn't train on user content and doesn't permit its providers to either. There's no voice cloning — only preset voices. A built-in assistant offers four help levels, from Co-Pilot down to none.

Is the price real?

The Starter tier costs $16 a month; Creator runs $35 and is free through the end of November. The bundle covers transcription, Headliner-style visualizers and Descript-style text editing — services Francis says run $20 to $30 each monthly. Users with existing ElevenLabs, Claude or OpenAI accounts can bring their own API keys and bill usage there.

Linnet stays in beta for now, partly so users can find bugs. Early feedback is positive, Francis said, though nobody has been charged yet — so the pricing has no revenue track record to check. His one regret is timing: he'd have shipped a year earlier, before the market "got saturated."