3D

Meng To released ThreeUI, an open-source library of procedural 3D landing-page components, icons, motion designs, and templates. Creators can copy its source or prompts into coding agents, while some extras remain paid.

SpAItial is recruiting creative partners for a beta that turns simple iPhone recordings into consistent multi-room 3D worlds. The company describes the system as a generative world model.

Spline rebuilt its web-based 3D editor with Agent Mode and an MCP connection for its desktop app. V2 also adds a WebGPU engine, PBR and HDR workflows, and scripting for code-enabled scenes.

Intangible connects to Claude Code, Claude Desktop, or ChatGPT through MCP so an agent can build scenes, compose cameras, and preview renders. The beta leaves creators to finish directing inside the scene.

Rainisto's workflow turns a screenplay into a rough Blender animation and uses it as Seedance video reference. The Blender timeline preserves planned cuts, lenses, and camera angles before generation.

Danny Limanseta says Grok 4.6 produced most of a WebGPU horde shooter, including its UI, visual effects, and sound effects. Meshy generated and rigged the character models separately.

Bilawal Sidhu demonstrates 2D moments reconstructed as 3D holograms and left at their original physical locations. One venue-scale spatial-memory project used more than 500 captures.

Meng To released the Kage landing page repo and reusable Three.js skills after building the 3D scrolling site with Claude Code, Opus 5, and image tools. He cites a 922 KB build and mobile fixes.

Meng To released code and prompts for a scrolling bookshelf demo refined through many Codex prompts. Karpathy separately used Claude Opus 5 to turn the opening of LOTR into a 5,500-line Three.js render.

Meng To used Opus 5 through Codex and Claude Code to rebuild Sakura Crossing as a single-file Three.js scene. Other demos show one-prompt Fallout-style RPGs and a Pool Rooms survival-horror level.

New Opus 5, Codex, and Fable demos showed agents creating 3D worlds, Blender assets, Godot ports, and one-shot games. Pushback focused on whether the clips become games worth playing.

Creators posted demos where Opus 5 built browser 3D games, Codex Sites produced a Three.js game, and Grok 4.5 helped redesign a Steam card game. Some workflows now pair video concepts with Claude Code, but the evidence is demo-based.

A demo roundup showed Grok 4.5 generating game engines, worlds, controllers, CAD objects, React projects, and voxel scenes. The clearest consequence was faster playable prototyping by a single operator.

Google DeepMind unveiled GenCeption, which turns video into depth, segmentation, camera rays, 3D keypoints, and prompt-steered scene representations. The thread links project and Hugging Face materials.

Demos from Gokay and Meng To used video, object-view, and image references to generate HTML, Three.js, or Blender scenes. The builds paired Sol or Claude Code with tools including Reve, HyperFrames, and Blender MCP.

Min Choi collected early GPT-5.6 Sol builds spanning Blender scenes, After Effects automation, Three.js characters, UI cloning, SQL DOOM, and Claude Code tests. Meng To also published a video-to-HTML workflow.

Min Choi collected Claude Fable 5 builds spanning a native Command & Conquer iOS port, a Subway Surfers clone, procedural 3D worlds, dungeon generation, and Blender assets. The examples show developers using Fable 5 across mobile ports, browser gameplay, and 3D asset workflows.

Tesana says Muranyi-3 turns plain-text prompts into playable 3D worlds with generated logic, NPCs, reference-art conditioning, and collaboration links. The launch also claims 90% fewer common errors than the prior model, but the evidence here comes from a promotional thread.

WWDC demos showed Apple Maps moving to radiance-field city scenes built from oblique aerial imagery, and follow-up posts say the rollout covers more than 300 cities. Posts also note better large-scale detail, but ground-level splats and a Vision Pro Maps app still appear absent, so users should watch for broader coverage.

Hyper3D launched Rodin Gen-2.5 with claims of 10 million-plus polygons, 4-second low-detail generations, and editable Bang-to-Parts separation. Try the same asset across Blender, Unity, Unreal, or 3D-print workflows and tune it from 4s to 80s.

SpAItial released Echo-2 HQ as a higher-detail version of its scene model and made it available in the app and API. Try it for closer zoom fidelity and direct handoff into Gaussian-splat style 3D workflows.

Magnific said 3D Scenes now uses Marble world generation to turn one reference image into a navigable 3D environment with lighting and depth intact. That creates a path from object photo to controllable scene and campaign-ready product imagery, so creators can skip a full 3D background build.

Google I/O demos showed Project Genie grounding explorable worlds in Street View and Maps imagery, with creators restyling real locations into playable scenes. The demos point to location-specific previs and game prototypes, though public access is still framed through event hands-ons.

Posts show an open-source toolkit that turns one reference image into an interactive 3D scene with generated meshes, lighting, physics, and sound. The demo stack chains World Labs, Hunyuan 3D, ElevenLabs, and fal rather than a single native model.

Stages AI demos show one-shot clips being turned into frames, prompts, storyboards, timelines, and Blender-ready scenes, with 100 camera rigs layered on top. The workflow compresses previsualization and 3D scene setup into one tool chain, though the evidence comes from a single creator and vendor account.

An 80.lv-cited report said Anthropic paid €240K to sponsor Blender, while Blender leadership said the deal was not an AI takeover and creator posts said the public patronage was later pulled. The connector still appears to ship, but the funding tie-in became a flashpoint inside open 3D communities.

Posts and demos described Claude Code controlling Blender through an MCP connector to create, edit, and debug 3D scenes from chat prompts. The demos matter because they show text prompting reaching direct scene manipulation inside a mainstream 3D tool.

SpAItial AI launched Echo-2, a physically grounded world model that turns one image into an interactive 3D scene and can distill meshes, point clouds, or 3DGS outputs. It matters for previs, environment design, and virtual production because creators get navigable scenes instead of only baked video clips, though the evidence is still mostly company-led demos.

A short workflow paired GPT Image 2 art with Freepik 3D Scenes to turn flat frames into explorable environments and adjustable camera angles. The result looks useful for previs and shot framing, but the demo stays at prototype-level geometry.

Posts introduced Modly as a fully local image-to-3D tool that turns one image into a mesh with drag-and-drop input and no cloud API. The release matters because 3D asset generation stays on-device, with current reporting concentrated in a single launch thread.

Tencent released HY-World 2.0 with WorldMirror 2.0 code and weights for turning text, images, or video into persistent 3D scenes. The output includes navigable geometry and camera data instead of disposable video frames.

NVIDIA introduced Lyra 2.0, which turns an image or walkthrough video into an explorable 3D environment exportable as Gaussians, meshes, and physics-ready assets. It shortens previs and worldbuilding from concept image to navigable scene.