Rendora launches text-to-3D video platform with editable avatars

Creating a decent 3D character video used to mean maxing out budgets on motion-capture suits, 3D animators, and endless rendering hours.

Rendora is trying to erase that entire pipeline. On August 9, fresh off demonstrations at SIGGRAPH 2024, the team rolled out its text-to-3D video platform on the web (rendora.ai), enabling anyone to whip up animated 3D avatar videos from plain text in seconds. At the Denver event, they even had attendees checking out the results inside an Apple Vision Pro.

Script to screen in seconds

The workflow feels more like directing a virtual stage than wrestling with animation software. Users pick an avatar and a 3D environment, paste in their script, and hit generate.

Within seconds, the platform spits out a fully lit scene with camera movement and voice-synced character animation. Under the hood, Rendora uses proprietary large language models to handle speech synthesis and body motion, keeping gestures and facial expressions aligned with the text.

The real kicker here is editability. Unlike diffusion-based video models where you are stuck with whatever random pixels get baked into the MP4, Rendora operates in actual 3D space. That means creators can tweak camera angles, swap avatars, and adjust layout elements after the initial generation instead of rerolling prompts from scratch.

Founder Jim Chai noted that the platform leverages millions of 3D assets and a real-time rendering engine, aiming to bring studio-grade production to standard marketing, corporate training, and digital content.

Why it matters

We think Rendora's biggest edge is choosing control over pure novelty.

Most generative video tools like Sora or Runway generate flat 2D video. They look incredible, but fixing a single bad frame or shifting the camera five degrees to the left is practically impossible. By marrying generative AI with real-time 3D graphics and virtual cameras, Rendora gives enterprise creators the precise control they actually need for professional output.

For small marketing teams, educators, and solo creators, this makes spinning up localized explainers and product pitches remarkably cheap. You no longer need to hire on-camera talent, book studio spaces, or manage render farms.

The challenge, of course, is avoiding the uncanny valley. If the asset library feels too repetitive or facial movements lack nuanced emotion, audience fatigue will set in fast. Still, packaging editable, real-time 3D generation into a browser app is exactly where virtual production needs to go.