← Plugin catalog
Creativity

Meshy 3D

Meshy LLC v1.1.1

From ideas to 3D creations

Create a free account to save this plugin.

From the developer

Declared capabilities

  • Text to 3D and Image to 3D Generation
  • AI Reference Images and Sketch to 3D
  • Game Props, Characters and Low-Poly Assets

Cost & access

This plugin does not sell products or process payments. Generating uses credits from the user's existing Meshy account, and downloads follow that account's plan.

Declared limitations

  • Rigging and animation need a terminal with the Meshy CLI. In ChatGPT without one, the plugin says they are not available here, installs nothing, and does not describe the model as rigged.
  • Print repair and multicolor 3MF need a terminal with the Meshy CLI. In ChatGPT without one, the plugin offers the STL download as the handoff and never calls the model printable.
  • No purchases or upgrades

Onboarding example

Create a low-poly treasure chest for my game with PBR textures.
Sources & dates

Developer declarations; not independently tested. Listing last checked Nov 14, 2026. Package observed Nov 14, 2026.

View package source →

Description

From the marketplace listing

Meshy 3D is an AI 3D model generator for ChatGPT and Codex. Turn text, photos or sketches into 3D assets for games, animation, AR and 3D printing, usually in about a minute. Create models with text to 3D and image to 3D, starting from an AI reference picture when the look is not settled yet. Make game-ready props and characters with low-poly topology and a target polygon count. Generate PBR textures from a prompt or a reference image, for new models and for models you already have. Follow progress and preview the result in the conversation, browse your model library in the Meshy 3D app, and download GLB, FBX, OBJ, USDZ or STL. In Codex and other hosts with a terminal, Meshy 3D goes further by running the Meshy CLI on your computer. Rig humanoid characters, add animations and generate motion clips from text. Remesh models, build LODs, unwrap UVs and convert formats. Prepare models for 3D printing with printability checks, mesh repair, splitting into parts and single-color or multicolor 3MF files, or explore Creative Lab to create figurines, lamps, keychains and fridge magnets. Hand print files to your slicer and save assets directly into your project. Requires a Meshy account; what you can generate and download depends on that account.

Language: English · Automatically detected from descriptions.

Matches for “code generation”

Exact text from the indicated source. A mention alone does not establish support for your task.

Publisher capabilities · listing

Text to 3D and Image to 3D Generation AI Reference Images and Sketch to 3D Game Props, Characters and Low-Poly Assets PBR Texturing for New and Existing Models In-Chat Model Preview and Model Library GLB, FBX, OBJ, USDZ and STL Downloads With a terminal: Humanoid Auto-Rigging, Animation and Text-to-Motion With a terminal: Remeshing, LODs, UV Unwrapping and Format Conversion With a terminal: Printability Analysis, Mesh Repair and Model Splitting With a terminal: Single-Color and Multicolor 3MF, Sizing and Slicer Handoff With a terminal: Creative Lab Figurines, Lamps, Keychains and Fridge Magnets With a terminal: Multi-View 3D Generation and Project Asset Delivery

Publisher full description

Meshy 3D is an AI 3D model generator for ChatGPT and Codex. Turn text, photos or sketches into 3D assets for games, animation, AR and 3D printing, usually in about a minute. Create models with text to 3D and image to 3D, starting from an AI reference picture when the look is not settled yet. Make game-ready props and characters with low-poly topology and a target polygon count. Generate PBR textures from a prompt or a reference image, for new models and for models you already have. Follow progress and preview the result in the conversation, browse your model library in the Meshy 3D app, and download GLB, FBX, OBJ, USDZ or STL. In Codex and other hosts with a terminal, Meshy 3D goes further by running the Meshy CLI on your computer. Rig humanoid characters, add animations and generate motion clips from text. Remesh models, build LODs, unwrap UVs and convert formats. Prepare models for 3D printing with printability checks, mesh repair, splitting into parts and single-color or multicolor 3MF files, or explore Creative Lab to create figurines, lamps, keychains and fridge magnets. Hand print files to your slicer and save assets directly into your project. Requires a Meshy account; what you can generate and download depends on that account.

What changed

· 8 observations

Capabilities & instructions

Supporting files updated; open the comparison for details. Skill instructions changed from “working CLI | A new print job runs on the CLI from its first step, as the sections below describe, whether or not Meshy's hosted tools are also listed: its...” to “terminal | A new print job runs on the CLI from its first...

View comparison →
Capabilities & instructions

Supporting files updated; open the comparison for details. Skill instructions changed from “`get_balance`) | Generation, reference pictures, texturing, lookups, downloads and the balance through the tools: [Tools](references/tools.md) | | The tools...” to “`get_balance`; a host may list them with a prefix, as in...

View comparison →
Capabilities & instructions

Full description changed from “Turn a description or a photo into a 3D model with Meshy, add PBR textures, and download it for games, AR and 3D printing.” to “Generate 3D models from text or images with Meshy, add PBR textures, and download them for games, animation, AR and 3D printing. In hosts with a...

View comparison →

Technical updates · 3

Package or technical metadata updated

View technical comparison →

Package or technical metadata updated

View technical comparison →

Package or technical metadata updated

View technical comparison →

Keywords

Files & skills

File archives

Plugin package24 files · 66.2 KBBrowse files →
Skill instructions
meshy-3d6.32 KB

View saved version →

---
name: meshy-3d
description: "Create 3D models with the Meshy 3D tools: a description or a picture to a 3D model, textures, reference pictures, and finding, showing and downloading the user's Meshy models."
license: MIT
---

# Meshy 3D

The Meshy 3D tools turn a description or a picture into a 3D model. This skill says how to use
them well. The tools enforce the user's plan, prices and limits: follow their results and error
texts, and never work around them.

## Pick the route

| The user has | Route |
|---|---|
| A description | `text_to_3d_preview` (untextured). Once it succeeded and they want color and materials, `text_to_3d_refine` with its `preview_task_id`. |
| A picture they uploaded | `image_to_3d` with `image_file` |
| A picture at a public link | `image_to_3d` with `image_url` |
| An idea whose look is not settled | `text_to_image`, show the picture, then `image_to_3d` with one of its `image_urls` as `image_url` |
| One of their untextured models (`texturable` is true) | `texture_model` with its `task_id` and a `texture_prompt` or `texture_image_url` |
| "My models", "the last one", "that fox" | `list_tasks` (`kind` models or images; `filter` untextured or textured), then `get_task` |

- Shape first: a preview is cheaper and faster than a textured model. Texture once the shape is
  right. `image_to_3d` textures by default; pass `should_texture: false` when only the shape is
  wanted.
- A prompt describes one object: what it is, its shape, material and style, in plain words, up
  to 800 characters. Turn a vague request into such a prompt and say what you sent.
- Game assets on a budget: `text_to_3d_preview` takes `target_polycount` (100 to 300,000) and
  `topology` (quad or triangle).
- Texture choices: `texture_resolution` is 2k by default; 4k and 8k only as the user's plan and
  the `ai_model` allow (`open_meshy_studio` returns them as `entitlements.texture`). `enable_pbr`
  adds PBR maps.

## Credits and consent

- `get_balance` returns the balance and `prices`: what each create costs with its default
  parameters, and in `prices.texture` what `texture_model` costs for each `ai_model` and
  `texture_resolution`. A call with other non-default parameters (another `ai_model`, 8k on
  `text_to_3d_refine`) costs a different amount that is not listed: say so instead of quoting a
  default price or guessing.
- Every create spends credits. The generation the user asked for proceeds. Before a paid step
  they did not ask for (texturing a preview, another variant, a higher resolution), say what it
  is and its price, and wait for their yes.
- When the balance is below the price, say so and do not start.
- `request_id`: a new random UUID for each new generation. A create answered `CREATING` or
  `RESULT_UNKNOWN` may still complete: resolve it with `get_task` on its `task_id`, or, when it
  came back without one, by repeating the same call with the same `request_id`. A `request_id`
  holds its create for 30 days: inside that time the repeat answers from the first call and
  creates nothing. After 30 days, or when you cannot tell when it was first sent, do not
  repeat it; look for the model with `list_tasks`, and if you cannot find it, ask before paying
  again. Never start it again with a new `request_id` unless the user agrees to pay twice.
- A refused create made no model and spent nothing; its error says so. Trying again, once the
  cause is dealt with, needs a new `request_id`.

## While it runs

- Right after a create, call `show_model` with the `task_id` and `task_type` it returned, where
  the host offers that tool. Where it shows Meshy's card, the card follows the progress by
  itself; where it does not, report the progress in text.
- Call `get_task` when you need the result to continue, for example before `text_to_3d_refine`.
  A model usually takes one to a few minutes; check at a calm pace, not in a tight loop.
- `FAILED`: give its `error`. Trying again is a new paid create, so ask first.

## Show and deliver

-

Preview of the first 4,000 characters. Read full instructions →

Referenced files: 2 — file references are metadata, not package contents.

meshy-3d-generation7.76 KB

View saved version →

---
name: meshy-3d-generation
description: "Generates and edits digital 3D assets with Meshy: text-to-3D and image-to-3D models, reference images, texturing, remeshing and LODs, UV unwrapping, format conversion (GLB, FBX, OBJ, USDZ, STL, 3MF, Blender), rigging, animation and motion clips from text. Uses Meshy's hosted tools first and the Meshy CLI for what they do not do. Use when the user asks to make, texture, rig, animate, convert, find or download a 3D model or game asset. For physical 3D printing use meshy-3d-printing."
license: MIT
compatibility: "Hosted tools: a host connected to Meshy's MCP server, nothing to install. CLI steps: a terminal with network access, Node.js 22.12+ and the Meshy CLI 0.8.x, 0.8.5 or later (npm install -g meshy-cli@0.8.5)."
---

# Meshy 3D Generation

Turn the user's request into a delivered asset. Completion means the requested files in the
requested location, or their download links where the host has no file system, and an available
preview inspected. Login or task submission alone is not completion. Choose the shortest
suitable pipeline and reuse existing assets for follow-ups.

## How to run it

Meshy is reached two ways: its hosted tools, and the Meshy CLI. **Use the hosted tools first;
the CLI is for what they do not do.** See which this session has before anything else.

| This session has | Run |
|---|---|
| Meshy's hosted tools (`image_to_3d`, `text_to_3d_preview`, `get_task`, `download_model`, `get_balance`; a host may list them with a prefix, as in `meshy_3d_image_to_3d`) | Generation, reference pictures, texturing, lookups, downloads and the balance through the tools: [Tools](references/tools.md) |
| The tools and a terminal | The same, plus the CLI for what the tools do not do: rigging, animation, motion, remeshing and LODs, UV unwrapping, multi-image-to-3d, conversion, resizing, and steps on a local model file. When the `meshy` command is missing, install it; that is the normal first run, not a reason to decline |
| The tools and no terminal | The tools only. For a step they do not do, say that it is not available in this environment |
| A terminal and no Meshy tools | Everything through the CLI: the CLI sections below and their references |

**The tools not offering a step never means Meshy cannot do it.** With a terminal, a step the
tools lack is a CLI step: run `meshy --version`, install the CLI when it is missing, and check
`auth status`, all as *CLI: first use and authorization* says. Do this before telling the user
that anything is unavailable.

The CLI counts as not working only after that was tried: the install was refused and the
temporary package failed too, or the terminal cannot reach Meshy (a sandbox without network
access, for one). Then say which of these happened, and that the step is not available in this
environment.

Choose the path when you plan the job, and say in the plan which it is:

- **The tools cover every step** (make, texture, find, show, download): the tools.
- **The request already names a CLI step** (a character to rig, an LOD chain, several photos of
  one object): set the CLI up first, then run the whole job on it. Its recipes set options at creation
  that the tools do not have, such as the pose of a character to rig, and its steps chain by
  task ID.
- **A CLI step comes up later**, on a model the tools made: the CLI continues from that model's
  task ID. Tools, *Continuing on the CLI*, says how, and how to check first that both are signed
  in to the same account.

## Read what this request needs

| Need | Reference |
|---|---|
| The hosted tools: routes, credits and consent, progress, delivery, continuing on the CLI | [Tools](references/tools.md) |
| CLI runner, credentials, or output location | [Setup](references/setup.md) |
| CLI generation and the models to pair for a project, texturing, remeshing, conversion, scale, or 2D images | [Pipelines](references/pipelines.md) |
| Rigging, animation, or generated motion (CLI) | [Animation](referenc

Preview of the first 4,000 characters. Read full instructions →

Referenced files: 8 — file references are metadata, not package contents.

meshy-3d-printing8.72 KB

View saved version →

---
name: meshy-3d-printing
description: "Generates and prepares models for physical 3D printing with Meshy: white and multicolor prints (3MF, STL, OBJ), printability analysis and repair, splitting into parts, Creative Lab products such as figures, lamps and keychains, sizing, and slicer handoff (Bambu Studio, OrcaSlicer, PrusaSlicer, Cura). Print preparation runs on the Meshy CLI; a host with only Meshy's hosted tools gets a model and its STL. Use when the user wants something 3D-printed or asks for print or slicer files."
license: MIT
compatibility: "Print preparation: a terminal with network access, Node.js 22.12+ and the Meshy CLI 0.8.x, 0.8.5 or later (npm install -g meshy-cli@0.8.5). Without one: a host connected to Meshy's MCP server."
---

# Meshy 3D Printing

Deliver the requested print files at the intended size, with analysis/repair results where
applicable and a slicer handoff when requested. This skill is self-contained. Ordinary digital
assets belong to meshy-3d-generation.

## How to run it

Print work is CLI work: sizing, analysis and repair, splitting, multicolor files, Creative Lab
and the slicer all need Meshy CLI **0.8** (0.8.5 or later). Meshy's hosted tools make and
download models; they do not prepare a print.

| This session has | Run |
|---|---|
| A terminal | A new print job runs on the CLI from its first step, as the sections below describe, whether or not Meshy's hosted tools are also listed: its recipes set output formats at creation that the tools do not have. When the `meshy` command is missing, install it; that is the normal first run, not a reason to decline. A model the tools already made continues on the CLI from its task ID, after the account check in [Tools](references/tools.md), *Continuing on the CLI* |
| Meshy's hosted tools (`image_to_3d`, `get_task`, `download_model`; a host may list them with a prefix, as in `meshy_3d_image_to_3d`) and no terminal | The model only: an untextured model from the photo or a reference image, and its STL download as the handoff ([Tools](references/tools.md)). Say that sizing, analysis, repair, multicolor files and the slicer are not available in this environment, and never call the model printable |

**The tools not offering a step never means Meshy cannot do it.** With a terminal, run
`meshy --version`, install the CLI when it is missing, and check `auth status`, all as *Cloud
authorization* says, before telling the user that anything is unavailable.

The CLI counts as not working only after that was tried: the install was refused and the
temporary package failed too, or the terminal cannot reach Meshy (a sandbox without network
access, for one). Then say which of these happened, and treat the session as the second row.

| Need | Reference |
|---|---|
| The hosted tools, and continuing on the CLI from a model they made | [Tools](references/tools.md) |
| CLI runner, credentials, or output location | [Setup](references/setup.md) |
| Local preparation, white or multicolor prints, analysis/repair, or splitting into parts | [Printing recipes](references/printing.md) |
| Creative Lab products (figures, lamps, keychains, magnets, keycaps, fidgets) | [Creative Lab](references/creative-lab.md) |
| Cost, waiting, preview, delivery, or finding a previous model | [Delivery](references/delivery.md) |
| Failed or uncertain operation | [Troubleshooting](references/troubleshooting.md) |

Read relevant sections as needed. Recipes illustrate valid commands; select only the stages
needed for the user's output.

## Choose local or cloud work

| Request | Route |
|---|---|
| Scale, orient or ground an existing OBJ | Local `mesh prepare-print`; no auth or balance |
| Open an existing STL or 3MF | Local `slicer detect` and `slicer open`; no auth or balance |
| White model from a photo, a sketch or text | Photo: image-to-3d. Otherwise recommend a reference image first (multi-view when depth or sides matter), with text-to-3d as the quicker alternative; geometry without textures; 

Preview of the first 4,000 characters. Read full instructions →

Referenced files: 8 — file references are metadata, not package contents.

Release notes

In hosts with a terminal, the skills now install the Meshy CLI on first use instead of calling its steps unavailable, upgrade a CLI that is too old for them, and tell you when a newer release is out. No change to the MCP server.

Declared in the saved package. Remote tools may change independently.

Package details

Publisher declarations from the archived package. These are separate from our research and the live service's terms.

Package license
MIT
Package author
Meshy LLC
Keywords
See keywords
Commerce declaration
Does not support commerceThis does not establish whether access is free or paid.
Review scenarios
5 positive · 3 negativeScenario counts, not test results.

Declared capabilities

  • Text to 3D and Image to 3D Generation
  • AI Reference Images and Sketch to 3D
  • Game Props, Characters and Low-Poly Assets
  • PBR Texturing for New and Existing Models
  • In-Chat Model Preview and Model Library
  • GLB, FBX, OBJ, USDZ and STL Downloads
  • With a terminal: Humanoid Auto-Rigging, Animation and Text-to-Motion
  • With a terminal: Remeshing, LODs, UV Unwrapping and Format Conversion
  • With a terminal: Printability Analysis, Mesh Repair and Model Splitting
  • With a terminal: Single-Color and Multicolor 3MF, Sizing and Slicer Handoff
  • With a terminal: Creative Lab Figurines, Lamps, Keychains and Fridge Magnets
  • With a terminal: Multi-View 3D Generation and Project Asset Delivery

Package observed Nov 14, 2026.

Technical details
First seen
Nov 3, 2026 · 06:00 UTC
Last seen
Nov 14, 2026 · 18:00 UTC
Latest observed change
Nov 13, 2026 · 12:12 UTC
Collection status
Collected

plugin_asdk_app_6ae4d5c0fba881918af7877ef49308f2

Download plugin data (JSON)

Plugin details: served from cache. Prepared . Refreshed when plugin data changes or after 6 hours. Your account and Follow status are live.