vork
Headless 3D and VFX engine built for AI agents to write scene code, shaders, and render directly on the GPU.
Programmatic GPU rendering engine designed for AI agents writing scene code, shaders, and visual effects.
vork is a headless 3D and VFX rendering engine designed specifically for autonomous AI agents. Rather than operating a visual user interface, agents construct scenes, materials, cameras, and custom shaders through code. It is aimed at developers building automated video pipelines, visual effects workflows, and programmatic 3D generation.
What it is
vork accepts scene data in JSON, Python, or TypeScript and compiles GPU programs to render full 3D environments and composited visual effects. The entire core API is context-optimized to fit inside approximately 2,500 tokens, enabling LLMs to iteratively write code, preview frames, and refine parameters. Rendering is driven by Rust and Slang on Vulkan or Metal backends.
Who it's for
It is ideal for software engineers, VFX technical directors, and AI developers building agentic video generation tools or automated 3D rendering pipelines. It is not intended for human artists looking for a traditional GUI-based 3D workspace like Blender or Unreal Engine.
Key features
Agent-Native API
Provides a compact 2,500-token API surface designed for LLM context windows to write and edit scene code.
Custom Shader Injection
Allows agents to write Slang compute shaders for light propagation, fluids, compositing, and custom materials.
Multi-Language Project Scaffolding
Generates ready-to-render project templates using plain JSON, TypeScript, or Python SDKs.
Vulkan and Metal Acceleration
Runs native GPU ray-tracing queries on Linux, Windows, and macOS.
Up and running in 3 steps
Our honest verdict
What's great
- Significantly faster frame rendering compared to traditional CPU/GPU Viewport engines like Blender EEVEE.
- Compact API footprint allows LLMs to manage complete scene code directly in context.
- Extensible GPU hooks enable custom physics simulations and procedural shader effects.
- Supports headless project initiation and API help inspection on machines without a GPU.
Where it falls short
- Rendering requires specific hardware such as Apple Silicon or an NVIDIA GPU supporting Vulkan ray queries.
- Lacks any visual GUI interface for manual asset placement or timeline manipulation.
- Currently released as a research preview with source-available licensing rather than standard OSI terms.
- Complex scenes require developer familiarity with CLI workflows, Slang shaders, or Rust execution.
Pricing & license
vork is source-available and free during its research preview phase. Offloading renders to Voyager cloud GPUs is available as an option if local hardware is insufficient.
Open remote AI-training roles
873 open remote AI-training roles, updated November 15, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Can I run vork on a machine without a dedicated GPU?
Scene authoring, code scaffolding, and CLI help commands run without a GPU, but local rendering requires Apple Silicon or an NVIDIA GPU with Vulkan ray queries.
Is vork a generative video diffusion model?
No, vork is a programmatic 3D ray-tracing engine and shader compiler, not an AI video generator.
