A new category of game engine, purpose-built for a world where NPCs talk, headsets are mainstream, and the browser is a first-class runtime. One API. Three targets. Ten thousand capabilities.
Unity, Unreal, and Godot were architected before ChatGPT, before the Quest Pro, and before WebGPU. RakuAI was designed from day one around all three.
Call the same REST endpoints from the browser, from a native Windows client, or from an XR headset. No separate WebGL port, no stripped mobile SKU, no platform-specific gotchas. Build once, ship to web, native, and immersive simultaneously.
LLM-driven NPCs, on-device small language model inference, emotion modeling, and dialogue trees are first-class subsystems — not plugins bolted on after the fact. The engine was designed around agents that can see, think, and speak.
18 native subsystems. 10,328 engine functions. 9,500+ REST endpoints. Physics, rendering, XR, audio, animation, networking, scripting — the full surface area of a real engine, not a framework or a demo. Battle-tested and ready for production workloads.
Existing engines were built for a world of single-player desktop games rendered in native binaries. RakuAI is built for a world where every character talks, every headset ships with hand tracking, and every game is one link away in a browser tab. The foundations had to be different.
| Capability | RakuAI | Traditional Engines |
|---|---|---|
| One API for web, native, and XR | Yes — single endpoint set | Partial — separate exports |
| LLM NPCs as a first-class subsystem | Built in | Third-party plugin |
| On-device small language model inference | Built in | Not available |
| OpenXR hand and eye tracking | Built in | Plugin-dependent |
| REST API for every capability | 9,500+ endpoints | Limited / none |
| Browser runtime with full feature parity | Yes | Stripped build |
| Patent-protected core technology | 5 granted US patents | N/A |
Browse the API reference, run the browser engine, or talk to us about enterprise deployments.
RakuAI is an NVIDIA Inception member. We are bringing NVIDIA-accelerated 3D Gaussian splatting to Raku Capture — a phone scan of any room turns into an explorable splat in minutes. Read more →
NVIDIA Inception is a free program for AI startups; RakuAI's membership does not imply that NVIDIA endorses RakuAI's products. Learn more at nvidia.com/startups.
RakuAI is built on AWS and an AWS Activate participant. Production reconstruction runs on Azure today; AWS GPU capacity (G5 / G6e) is on our roadmap, honestly gated until quota lands.
“AWS” and “AWS Activate” are trademarks of Amazon.com, Inc. or its affiliates. RakuAI’s participation does not imply that AWS endorses RakuAI’s products.
RakuAI is an AWS Activate participant. Raku Capture runs on AWS — GPU training, splat storage, and backend compute are powered by AWS infrastructure. Read more →
AWS Activate is a program for startups; participation does not imply that Amazon Web Services endorses RakuAI’s products. Learn more at aws.amazon.com/activate.