MediaPipe Face Landmark is a real-time facial tracking AI that maps 468 facial points to animate 3D models using morph targets. Designed for developers, it enables applications like AR filters, avatar animations, and emotion detection by translating live camera input into precise facial movements. The technology powers web-based demos (like Bolt's mascot project) where expressions drive 3D model deformations seamlessly.

Unlike static face filters, this tool offers low-latency tracking with web compatibility (via Three.js) and integrates with MediaPipe’s broader computer vision suite. Ideal for creators building interactive experiences, it balances accessibility—through open-source examples—with professional-grade accuracy for gaming, virtual meetings, or accessibility tools.
Key differentiators:

Real-time 468-point facial mapping

Morph target integration for 3D animation

Web-first (Three.js) implementation

Part of Google’s MediaPipe ecosystem

Open-source with ready-to-clone demos

MediaPipe Face Landmark

LegoGPT is a text-to-3D generation tool that creates physically stable LEGO brick models from written prompts. Trained on a dataset of over 47,000 stable LEGO structures, the model generates designs using 1-unit-tall cuboid bricks within a 20x20x20 grid. It ensures each model is buildable by integrating physics-aware constraints during generation.

Designed for a fixed set of 21 object categories, LegoGPT produces functional and visually coherent LEGO structures based on user input. The tool emphasizes physical feasibility, making the outputs suitable for manual or robotic assembly. It also introduces a dataset, StableText2Lego, and supports future development in AI-guided 3D design.

LegoGPT

Skybox AI lets you generate breathtaking 360° worlds instantly using just text prompts—no design skills needed. Whether you want a futuristic cityscape, a fantasy castle, or an underwater paradise, describe your vision and the AI brings it to life. You can then remix or edit these environments, adding or removing elements to perfect your virtual world.

Ideal for game developers, VR creators, and digital artists, Skybox AI turns imagination into immersive 3D spaces with a single click. Its library of presets (like Neo Tokyo or Aurora Nights) jumpstarts creativity, while customization options ensure every scene matches your unique vision.

Skybox AI

Spell AI transforms simple images into fully realized 3D worlds in minutes using advanced techniques like Gaussian Splatting and NeRFs. It generates consistent, high-quality scenes that can be rendered and explored, making it perfect for creating immersive environments or prototyping designs. This cutting-edge tool is currently in early access, offering a glimpse into the future of 3D content creation.

Ideal for developers, designers, and researchers, Spell AI simplifies the process of building detailed 3D worlds. Whether you're working on creative projects or exploring new technologies, it provides a fast and innovative way to bring your visions to life with stunning accuracy and depth.

Spell by Spline

Other 3D Tools