Content & Design

3D2cut SA offers comprehensive digital vine pruning training solutions designed to improve vineyard health and productivity. Co-founded with Simonit & Sirch, the platform provides short video lessons and interactive exercises, including pruning cut simulations, to teach various pruning methods in multiple languages. It also features manager dashboards for tracking progress and an innovative AI/AR pruning guidance system, which uses augmented reality glasses to suggest optimal cut zones. This tool addresses challenges like inconsistent pruning quality, high training burdens for new crews, and the increasing complexity of modern viticulture, making expert knowledge accessible and repeatable.

sd-webui-deforum is an official extension for AUTOMATIC1111's Stable Diffusion webui, designed to facilitate the creation of AI-generated animations and videos. Users can install it by cloning the repository into their extensions directory, downloading it manually, or installing it directly from the Extensions tab within A1111. The tool features a dedicated Deforum tab for entering animation settings, including prompt keyframing functions. It supports both 2D and 3D animation modes, with considerations for VRAM usage, especially in 3D mode. After generation, users can view the video or GIF result directly within the GUI. The project also provides resources like a Discord community for support and sharing creations, and an Issues tab for reporting problems.

Menzo by CortexUI provides a comprehensive digital menu platform for restaurants, enabling them to create interactive menus with QR codes, 3D/AR models, and WhatsApp ordering. Restaurants can manage multiple venues from a single account, with support for up to 500 dishes per menu. The platform offers five customizable templates (Modern, Classic, Minimal, Premium, Dark) to match branding. Beyond digital menus, Menzo includes features for point-of-sale (POS) operations, inventory management, and staff management, making it a complete solution for restaurant operations. Real-time updates ensure that menu changes are instantly reflected for customers, eliminating the need for reprinting physical menus.

PixFlow Motion Magic ✨ is an AI-powered tool designed to transform private images into high-quality animated adult videos. This platform allows users to upload their own images and then customize various aspects of the resulting animation, including the overall style, the duration of the video, and the background music. The tool aims to provide a personalized and creative experience for generating dynamic video content. It is hosted on Hugging Face Spaces, indicating its accessibility and potential for community-driven development or usage. The focus on animated adult videos suggests a niche application for content creators interested in this specific genre.

ThreeDPoseUnityBarracuda is an open-source Unity sample project designed for 3D pose estimation, leveraging the Barracuda neural network inference library. This tool allows developers to implement real-time motion capture, enabling an avatar (like Unity-chan) to mimic human movements from a video input. It supports loading ONNX models for improved accuracy and provides options for choosing target videos, avatars, and even using a web camera for input. While the project is not actively maintained, it serves as a valuable foundation for integrating advanced pose estimation capabilities into Unity-based game development and other interactive applications. Users can customize avatar sizes and input sources, making it a flexible starting point for various motion-related projects.

Volinga Suite is an advanced tool designed for creators to effortlessly create and render Radiance Fields in real-time using Unreal Engine. The suite comprises three key components: Volinga Renderer, Volinga Exporter, and Volinga Creator. Volinga Renderer allows real-time rendering of Volumetric Radiance Field data, powered by the NVOL file format, and integrates as a plugin for Unreal Engine. The Volinga Exporter converts .ckpt files from NeRFStudio and .ply files from 3D Gaussian Splatting into NVOL files. Volinga Creator offers a user-friendly interface to streamline the process of creating NeRF models from images or videos, with options for online or local training. It also supports integration with Disguise RenderStream, Pixotope, and Nuke Server for professional workflows.

SurroundOcc is an advanced AI tool developed for multi-camera 3D occupancy prediction, primarily targeting autonomous driving applications. It reconstructs comprehensive and consistent 3D scenes by extracting multi-scale features from camera images and lifting them to 3D volume space using spatial cross-attention. The tool then applies 3D convolutions for progressive upsampling and multi-level supervision. A key differentiator is its pipeline for generating dense occupancy ground truth from sparse LiDAR points, leveraging existing 3D detection and semantic segmentation labels without requiring extra human annotations. This process fuses multi-frame LiDAR points for dynamic objects and static scenes separately, followed by Poisson Reconstruction and voxelization to create dense volumetric occupancy. SurroundOcc supports both occupancy prediction and ground truth generation on custom data, offering flexibility for researchers and developers in the autonomous driving domain.

MOVE Ai pioneers and perfects markerless motion capture systems, enabling high-fidelity 3D animation directly from video. Since 2019, the company has developed multi-camera systems and patented AI technology for its award-winning motion engine. This technology dramatically reduces production costs by eliminating the need for suits or markers, leading to faster shoot times and scalable volumes. It provides comparable data quality to optical motion capture systems, making it a valuable tool for leading studios in VFX, entertainment, and gaming. MOVE Ai aims to streamline the animation workflow and make motion capture more accessible and efficient for various creative projects.

V3D is an open-source implementation of the research paper "V3D: Video Diffusion Models are Effective 3D Generators." This tool leverages video diffusion models to create 3D content, offering capabilities such as generating dense multi-views from a single image and reconstructing 3D assets using techniques like 3D Gaussian Splatting or NeuS. It provides instructions for installation, downloading weights, and running scripts to generate and reconstruct 3D models. The project is actively being developed, with plans for more checkpoints and examples, making it a valuable resource for researchers and developers interested in advanced 3D generation from video data.

Zero123++ Demo Space is an AI tool hosted on Hugging Face, designed for generating 3D models. It provides a platform for users to explore and experiment with the capabilities of the Zero123++ model in creating three-dimensional assets. While the tool aims to offer a hands-on experience for 3D model generation, the live website indicates that it is currently experiencing runtime errors and scheduling failures, making it temporarily unavailable for use. Despite these technical issues, its presence on Hugging Face suggests it is intended to be a free resource for the community to engage with AI-powered 3D creation.

HoLa-BRep is an AI-powered tool designed for generating 3D CAD models in boundary representation (BRep). It offers a versatile approach to 3D model creation, allowing users to input data from diverse sources such as point clouds, sketches, text descriptions, or even single-view images. Upon processing, the application generates four plausible 3D models, providing users with multiple options to choose from. This capability makes HoLa-BRep a valuable resource for professionals and enthusiasts in fields requiring rapid prototyping, design iteration, or conceptualization of 3D objects from various starting points. The tool aims to streamline the initial stages of 3D modeling by automating the generation process.

AI Reelity is an innovative AI-powered travel planning tool designed to help users explore cities like a local while still experiencing popular tourist attractions. It generates tailored travel plans by offering dual perspectives: authentic local insights and top tourist hotspots. The platform features an intuitive interface, allowing users to simply enter a city and receive a personalized guide in minutes. AI Reelity adapts to individual tastes and interests, crafting unique experiences from iconic landmarks to hidden local gems. It also includes a "Movie Tourism" feature to discover filming locations and a "Tourist vs Local Trip Planner" for a comprehensive city exploration. The tool aims to save users time on planning, enabling them to spend more time experiencing their destination.

HyperLandmark is a free and open-source tool designed for real-time face landmark detection, primarily targeting mobile applications. It utilizes deep learning to accurately identify 106 facial landmark points, offering a detailed facial contour description. The tool is noted for its high accuracy, even in challenging lighting conditions, and its efficient, small model size (around 2MB for the tracking model), making it highly suitable for mobile integration. It also supports multi-face tracking and boasts fast processing speeds, with the Android version achieving 7ms per single face on a Qualcomm 820. The project provides both Android and Windows implementations, with the Android version based on deep learning and the Windows version on traditional SDM algorithms.

MeshAnything is a powerful 3D & Animation tool hosted on Hugging Face Spaces, designed to streamline the process of creating low-poly, artist-crafted meshes from existing 3D models. Users can upload OBJ files, which the application then normalizes and processes. The core of the tool is a transformer that intelligently generates optimized meshes. For enhanced flexibility, MeshAnything offers an optional marching-cubes preprocessing step, allowing for different approaches to mesh generation. This tool is ideal for artists and developers looking to efficiently convert high-detail models into more manageable, game-ready or stylized assets.

MyArchitectAI is an AI rendering software designed for architects and interior designers, enabling the creation of photorealistic architectural and interior renders in under 10 seconds. The platform is browser-based, eliminating the need for software installations and making it compatible with any device. It supports designs from various CAD and 3D modeling software, including SketchUp, Archicad, Revit, Rhino, Chief Architect, and Vectorworks, by accepting JPG/PNG formats. Key features include AI style transfer for experimenting with design variations, an AI render editor for quickly modifying materials and objects, and an AI enhancer for improving low-resolution renders. Users can also generate cinematic animations from still renders in 90-120 seconds, choosing from various camera presets.

Real3DPortrait is an open-source project providing a PyTorch implementation for one-shot realistic 3D talking portrait synthesis. It allows users to generate high-quality talking face videos from a single source image and a driving audio or video. The tool supports both audio-driven and video-driven methods for generating expressive 3D portraits. Key features include the ability to control mouth amplitude, map initial poses, and provide custom background images. It offers a command-line interface, a Gradio WebUI, and a Google Colab notebook for inference, making it accessible for various users. The project also provides training code for its audio-to-motion and image-to-plane models.

AI Motion Control is a premier platform designed for professional AI Motion Control, enabling creators to transfer complex movements and facial expressions from reference videos to static images with high precision. The platform utilizes advanced AI algorithms to ensure full-body synchronization, maintaining character identity across frames without distortion. It supports motion retargeting for various character styles and intelligently adjusts motion to fit target character proportions. Beyond movement, it clones nuanced facial expressions, including lip-sync and gaze tracking, for emotional storytelling. The tool also offers an integrated action library, advanced temporal consistency for smooth, flicker-free long-form videos, and style transfer persistence to maintain artistic integrity. All processing is handled via high-performance cloud rendering, eliminating hardware constraints.

SpatialLM is a 3D large language model designed to process 3D point cloud data and generate structured 3D scene understanding outputs. It can identify architectural elements such as walls, doors, and windows, as well as oriented object bounding boxes with their semantic categories. A key differentiator is its ability to handle point clouds from diverse sources, including monocular video sequences, RGBD images, and LiDAR sensors, unlike previous methods that often required specialized equipment. This multimodal architecture bridges the gap between unstructured 3D geometric data and structured 3D representations, providing high-level semantic understanding. SpatialLM enhances spatial reasoning capabilities for applications in embodied robotics, autonomous navigation, and other complex 3D scene analysis tasks. It offers models like SpatialLM1.1-Llama-1B and SpatialLM1.1-Qwen-0.5B, available on Hugging Face, and supports detection with user-specified categories.

DreamTalk is an open-source framework designed for generating expressive talking head videos. It utilizes diffusion probabilistic models to create high-quality videos that capture diverse speaking styles. The tool is robust, handling a wide array of inputs including songs, speech in multiple languages, and even noisy audio, and can work with out-of-domain portraits. Users can specify audio paths, style clips, head poses, and input images to generate videos. While the primary focus is on accurate lip-sync and vivid expressions, the resolution can be improved using external solutions like CodeFormer or MetaPortrait's Temporal Super-Resolution Model. The project provides inference code and pretrained checkpoints, though access to checkpoints requires an email request for academic research purposes.

Nak3D offers a comprehensive 3D asset infrastructure layer, focusing on provenance, rights management, and visualization. It leverages AI for 3D generation, converting photos into production-ready 3D assets with human QA. A key differentiator is its invisible fingerprinting technology, embedding proprietary ownership data at the pixel level, which survives file format conversions and compression. Every asset also receives a permanent, tamper-evident timestamp, aligning with EU DPP frameworks. Nak3D provides three product lines: LOCK for individuals to document valuable possessions, LITE for small businesses needing 3D product imagery, and DISCO for enterprises requiring full-service 3D asset creation and rights management. The platform aims to build a verified, fingerprinted 3D asset library, with customer-consented assets entering a vault for potential licensing.

AnySplat is an open-source tool designed for feed-forward 3D Gaussian Splatting from unconstrained views. It utilizes a transformer-based geometry encoder followed by three decoder heads to predict Gaussian parameters, depth maps, and camera poses. These outputs are then used to construct pixel-wise 3D Gaussians, which are voxelized and rendered into multi-view images and depth maps. The tool supports training and inference, with code available for installation and quick start. It also includes a Gradio-based demo for visualizing reconstructed 3D Gaussian Splats from uploaded images or videos, making it a valuable resource for researchers and developers in computer vision and graphics.

bullet3 is the official C++ source code repository for the Bullet Physics SDK, offering real-time collision detection and multi-physics simulation capabilities. It is widely used across various domains including virtual reality, game development, visual effects, robotics, and machine learning. The SDK supports a range of platforms like Windows, Linux, Mac OSX, iOS, and Android, and includes experimental OpenCL GPGPU support for accelerating collision detection and rigid body dynamics. Users can also leverage PyBullet, Python bindings for enhanced support in robotics, reinforcement learning, and VR, with simple installation via pip. The project is licensed under the permissive zlib license.

PrometheanAI is an AI assistant designed for professional creative teams involved in building virtual 3D worlds. It functions as an AI engine that understands various creative assets, including images, videos, 3D models, 3D animations, PDFs, and PPTs, reasoning about them to configure novel combinations for content creation. The tool aims to significantly speed up digital art production by handling mundane tasks, allowing artists to focus on creativity. A key differentiator is that it does not require users to upload their assets to the cloud or change their existing 3D editors, ensuring data privacy and seamless integration into current workflows. PrometheanAI supports major engines like Unreal Engine, Unity, 3ds Max, Maya, and Blender, offering open-source plugins for customization.

Wordcraft3D is a tool designed to help users generate 3D models from text prompts. It provides downloadable files in the .obj format, making it compatible with various 3D software. The platform offers a free trial, allowing users to experiment with its capabilities and view examples of generated models, such as apples, giraffes, and pineapples. However, the website currently indicates that the tool is experiencing connectivity issues with its backend server and is not functional. The developer suggests Meshy.ai as an alternative while they address the problem.