Company profile · 3 min read
Dollhouse: interactive AI shows from a single sentence
Dollhouse is building AI-generated interactive shows in which creators shape scripts, characters, voices and scenes while viewers influence the story.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
Dollhouse is building interactive AI shows where a single prompt becomes the screenplay, characters, voices and video, then the viewer changes where the story goes. Its likely buyers are creators, studios and apps looking for a new format between scripted video, games and character companions.
What it does
The South Park Commons profile describes interactive shows generated from one sentence, with viewers shaping storylines and character journeys. The current Dollhouse site frames the product as two sides: writers create worlds, refine scripts and direct scenes; viewers discover original shows and influence what happens next. The current site is recruiting a waitlist rather than publishing pricing or a public show catalog.
The founder context is compact but relevant. Divanny Lamas describes the company as building real-time avatar and companion infrastructure for creators, studios and apps; Michael Kass is listed as co-founder. The public pages do not provide a current production timeline, content-safety policy or rights model for generated characters, voices and stories.
Why I’d look closer
The advantage is an interaction loop that gives viewers agency without requiring a full game engine. A creator could test a world, character or branching scene quickly, while the audience supplies the next story beat through participation.
The tradeoff is coherence and ownership. Generated voices, characters and scripts raise likeness, copyright, moderation and child-safety questions. A branching show also needs a continuity model: if every viewer changes the plot, what remains canonical and what does the creator actually control?
What I’d ask
Who owns generated scripts, characters, voices and viewer branches? How are likeness, music and training rights handled? What controls exist for violent, sexual or manipulative content? Can a creator approve a scene before publication and export the underlying assets? How are audiences protected when AI characters feel socially real?
My editorial take
Keep Dollhouse on the shortlist for interactive-story pilots, not as a finished entertainment platform. The first useful proof is a short show with clear creator controls, branching continuity and a rights-safe asset pipeline. The idea is legible; the product decision is whether audience agency adds more than it costs in moderation and story coherence.
Quick facts
| Field | Sourced detail |
|---|---|
| Product | AI-generated interactive shows with scripts, characters, voices and video |
| Buyer | Writers, creators, studios and apps |
| Availability | Waitlist on the checked homepage; public pricing/catalog not published |
| Key unknown | Rights, moderation and continuity model for viewer-shaped stories |
| Main question | Can interactive generation preserve creator control and audience safety? |
Sources checked
| Source | Checked |
|---|---|
| South Park Commons Dollhouse profile | 2026-09-19 |
| Dollhouse homepage | 2026-09-19 |
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.
First1000 ↗ · X ↗