Myspace founders turn your selfies into AI video

Myspace's original architects are staging a comeback, and this time they want to put your actual face inside an AI-generated scene.

Plai Labs, an a16z-backed startup founded by Myspace co-founders Chris DeWolfe and Aber Whitcomb, just rolled out a free text-to-video generator called PlaiDay. Instead of just creating abstract clips from prompts, it takes your selfie and sticks an animated version of you directly into the action.

Three seconds of fame, with some uncanny quirks

If you have ever wanted to see yourself flying across a canyon or battling aliens, the pitch here is simple: drop in a selfie, write a prompt, and get a short clip back to post online.

Right now, those clips are capped at just three seconds, and they come without audio, though the team says longer durations and sound are coming.

In testing, results are a mixed bag. An official demo showing an 1800s London street handled background movement and close-up facial expressions reasonably well. But when personalizing with selfies, things can get weird. One test without facial hair resulted in patchy textures fading into the skin, while some user creations on Discord ended up genuinely distorted.

Plai Labs suggests steering the model by adding terms like "close up" or "portrait of" to prompts. For now, the system handles single faces far better than crowd scenes.

Anyone can try it today inside PlaiDay's Discord server. The feature is scheduled to land in an alpha mobile app that also promises TikTok integrations, style transfer, and automated captioning.

The enterprise engine behind the social toy

While personalizing memes is what grabs attention, Plai Labs is building toward a much bigger infrastructure play.

The tool is powered by Orchestra, an internal AI platform that stitches together proprietary and open-source models. Jim Benedetto, former Myspace tech VP and Plai Labs co-founder, claims the system uses reusable building blocks that let a single product manager build capabilities that would normally require a team of five to ten PhDs.

According to Whitcomb, Orchestra is intended to handle far more than video avatars. The goal is to power automated marketing pipelines, image analysis, and security tasks. That broader platform is set to launch publicly in the coming months.

Why it matters

PlaiDay is taking an end-run around the cinematic video race by tapping into the oldest trick in social media: digital vanity.

While rivals push for hyper-realistic physics and longer shots, we think focusing on personal likeness is a clever shortcut to virality. Everyday consumers care far more about seeing themselves do something absurd than about generating yet another photorealistic drone shot of a coastline. For quick social gags, a three-second clip gets the job done without any technical learning curve.

However, the concept faces serious shelf-life hurdles. Personalized avatar trends flare up fast and burn out even faster. If the generation quality remains glitchy and prone to nightmare-fuel distortions, people will move on once the novelty wears off.

More importantly, using a consumer face-swapping demo to validate an enterprise AI orchestration platform is an awkward bridge. Whether businesses will trust a platform best known for cartoonish selfie clips to handle their marketing automation and analytics remains an open question.