Changelog

See all of the latest features and updates to the Meshy webapp.

September 2026

3D Model Generation

  • Released Meshy 7.1 for Image to 3D, with more geometric detail than Meshy 7. Ultra Mode is replaced by three geometry resolution options: Standard, Ultra 2K and the new Ultra 4K. Ultra 2K and Ultra 4K are available on paid plans, and Ultra 4K works with single-image input.

3D Model Post-Processing

  • OBJ, STL and resized GLB downloads are now significantly faster, and files with non-English names no longer fail to download.
  • Auto Split now takes your direction. Just describe the parts you want to separate, a dragon's wings or legs, a toy's head or boots, a car's wheels, and Meshy splits it your way, with more control over what comes apart.

3D Agent

  • The 3D Agent is now available on the Free plan.
  • The 3D Agent now runs on a more capable foundation model, delivering significantly better performance and more capable 3D creation.
  • Templates: the shortcut to 3D creation. 21 official templates now live on Meshy. Upload a photo, hit generate, and get a 3D model worth showing off. Browse them in Discover.

Teams & Management

  • Credit limits for team members are now manageable in the webapp, giving teams more control over credit usage when generating assets with the webapp and 3D Agent. This is separate from API key credit limit management.

August 2026

3D Model Generation

  • Released new Meshy 7 model for Image to 3D with Ultra Mode for finer surface detail, and Meshy 7 Texture, with sharper surface detail and cleaner PBR materials than Meshy 6. Texturing supports multi-view input from up to four ordered views of the same object. Both are available in Workspace and the 3D Agent.
  • Added Meshy 6 Lite, a faster, lighter variant of Meshy 6. It replaces Meshy 5 in the model picker without affecting existing Meshy 5 models and is now the fastest option for both model generation and texturing.
  • Smart Topology now generates in approximately 2 seconds, down from around 15.

3D Model Post-Processing

  • Added Semantic Segmentation, which splits models along semantic parts such as head, limbs and torso rather than geometry alone. Auto Mortise-and-Tenon generates interlocking connectors on the resulting parts, so printed pieces assemble without glue.
  • The online viewer now opens IGES, IGS, IFC, VRM and G-code. File Converter adds 3DS, AMF and VRM as source formats, and FBX as an output format.

3D Agent

  • The 3D Agent now asks before spending credits: approve each action, allow all actions until the task completes, or allowlist specific tools, with a per-turn credit budget. New users start in approval mode. Added a credit usage breakdown panel in settings, and each model's cost now appears at selection.
  • Added Skills: Meshy-curated presets triggered from the composer, with a showcase on the 3D Agent home. Logged-out visitors can browse trending skills.
  • Redesigned the 3D Agent home for improved user experience.
  • Semantic splitting and AI Auto-Repair are now available in the 3D Agent.
  • Asset cards now carry print and convert actions, remesh is available in the canvas with the same controls as Workspace, and images have their own toolbar. Assets can be downloaded in batches from the gallery and canvas, and any Workspace asset can be opened directly in the 3D Agent.
  • Model generations in the 3D Agent can now be retried for free. The number of free retries per generation depends on the subscription plan.

Creative Lab

  • New category: Pet Keepsake
  • Chibi Figure now supports pets and multi-person subjects.
  • Keycap build and colorize are now a single step. Builds save to the asset list, and head pose and size persist between sessions.

July 2026

3D Model Generation

  • Added Smart Topology, a new AI model for Image to 3D that generates low-poly models with native part segmentation and controllable polygon count.
  • Texture generation now supports 8K resolution for base color maps, with a new multiview diffusion algorithm that captures finer detail on realistic textures and removes lighting more cleanly.
  • Retired Prompt Helper. Prompt writing help is now handled by the 3D Agent.

3D Model Post-Processing

  • Auto Split is now available for 3D printing. It splits a model into watertight, printable parts in about 40 seconds. Parts can be auto-arranged on the build plate and exported to your slicer in one click.
  • Added Text to Motion to generate animations for your rigged character from a text prompt. Results can be saved to your motion library.
  • Improved Multi-Color Print with AI Optimize, reducing excess colors on complex models for better print results.

3D Agent

  • Added an Inpaint tool that repaints brush-selected areas of an image from a text prompt.
  • Image backgrounds can now be removed in one click.

Creative Lab

  • Added two new categories: Plant Pot and Tabletop Terrain.

June 2026

3D Agent

Released Meshy 3D Agent (Beta), the world's first AI agent for 3D creation. The Beta covers:

  • End-to-end creation in one conversation: brainstorming, batch visual concept generation, model generation with 4K textures, and export to DCC tools or slicers.
  • Web search for reference images, with results directly usable in generation.
  • A built-in knowledge base that answers your 3D creation questions.
  • Read-only conversation sharing via link, no login required.

Creative Lab

  • Added two new categories: Twist Egg and Golden Boot Keychain.
  • Creations are now automatically saved to the asset panel and can be resumed for further editing.

DCC Integrations

  • Added a ZBrush plugin that sends models from the Meshy workspace to ZBrush in one click as clean OBJ geometry.