Features Meshy 7 3D Agent API Pricing FAQ Get API Key
◆ Meshy APINow on Meshy 7.1

Build 3D Generation Into Your Product

A REST interface to the same generation engine that powers the Meshy web app 23 endpoints for text-to-3D, image-to-3D, AI texturing, rigging, animation and 3D printing prep, running on Meshy 7.1's Ultra 4K pipeline.

Get API Key → Read the Docs ↗

99.9% uptime SLA · ISO 27001, SOC 2 & GDPR compliant · Inputs never used for training · Pro plan+ required

Meshy API console dashboard showing usage, success rates and credit consumption


Latest on the API

What's new

Sep 10, 2026

Meshy 7.1 Ultra 4K via API

Text to 3D and Image to 3D now default to the Meshy 7.1 engine, with Ultra 2K/4K texture tiers callable the same way as standard generation. See Meshy 7 benchmarks.

Aug 12, 2026

Meshy 7 geometry engine

The default model behind every Text/Image to 3D call rebuilt alignment-in-training-loop preferred 63.8% of the time over Tripo 3.1 in blind testing.

Jun 4, 2026

Agent Skills & MCP server

The API is now the backbone of the 3D Agent ecosystem: open-source Agent Skills and an MCP server let Claude Code, Cursor and other assistants call these same endpoints directly.

Jul 2026

3D Print suite endpoints

Printability analysis (free), auto-repair and Auto Split are now callable via API for print-pipeline automation. See 3D printing on Meshy.

2026

Smart Topology on Meshy 7

Retopology-aware output that's both animation-ready and 3D-print-ready from the same generation call, no separate remesh step required.

2026

Legacy models still callable

Meshy-5/6 remain available at roughly 75% lower credit cost than Meshy 7 for draft-quality iteration.



23 endpoints

Every endpoint and its credit cost

Async task pattern throughout: a POST request creates a task and returns an ID immediately; poll for status or register a webhook to get the result when generation completes.

EndpointCategoryCredit Cost
Text to 3D (preview/mesh)Generation20 credits
Text to 3D (refine/texture)Generation10 credits
Image to 3DGeneration20-30 credits
Multi-Image to 3DGeneration20-30 credits
AI Texturing 2KTexturing10 credits
AI Texturing 4KTexturing10 credits
AI Texturing 8KTexturing15 credits
RetextureTexturing10 credits
RemeshProcessing5 credits
Auto-RiggingCharacter5 credits
AnimationCharacter3 credits
Text to Image2D3-9 credits
Image to Image2D3-12 credits
Convert file typeProcessing1 credit
ResizeProcessing1 credit
Analyze Printability3D PrintingFree
Repair Printability3D Printing10 credits
White Model Print3D Printing20 credits
Multicolor Print3D Printing40 credits
Creative Lab (Keychain/Magnet/Figure)Creative6 cr prototype → 20-30 cr build
Creative Lab (Lamp)Creative6 cr prototype → 30 cr build
Balance checkUtilityFree
WebhooksUtilityFree

Full up-to-date pricing lives on the official API pricing page pay-before-you-go, separate from web subscription credits, with volume pricing available through sales.



Quickstart

Authentication & your first call

API keys follow the format msy-<random-string> and authenticate via a Bearer token. A free test-mode key returns zero-cost mock responses so you can build against the API before spending credits.

Python
import requests headers = {"Authorization": "Bearer msy-YOUR_API_KEY"} # 1. Create a Text to 3D preview task resp = requests.post( "https://api.meshy.ai/openapi/v2/text-to-3d", headers=headers, json={"mode": "preview", "prompt": "a low-poly fox statue"} ) task_id = resp.json()["result"] # 2. Poll until the task succeeds status = requests.get( f"https://api.meshy.ai/openapi/v2/text-to-3d/{task_id}", headers=headers ).json()

Free test mode

Use the dummy key below to build your integration without spending real credits every response is a zero-cost mock.

msy_dummy_api_key_for_test_mode_12345678

Commercial use

Paid Meshy users own their generated assets outright. See commercial use and license terms.



Rate limits

By plan tier

Exceeding your tier's limit returns a 429 Too Many Requests response. See plans to upgrade.

TierRequests/secMax Queued Tasks
Pro2010
Premium2030
Ultra20100
Studio2060
Enterprise10050 (customizable)

Asset retention: 3 days on self-serve plans, indefinite on Enterprise. See API pricing for volume discounts.



Built-in tooling

Playground, dashboard & logs

Test endpoints live in-browser before writing any code, then monitor success rates, latency and credit consumption per request once you're in production.



Plug it in

Plugins, CLI & the MCP server

No installation required for the REST API itself it's cloud-based. On top of it, Meshy ships ready-made integrations, including the same Agent Skills and MCP server that power chat-based 3D creation in coding assistants.

🧊 Blender Plugin 🎮 Unity Plugin 🎮 Unreal Engine 🎮 Godot 🎬 Maya 🎬 3ds Max 🌐 Omniverse 🧱 Roblox ⌨️ CLI Tool 🔌 MCP Server (Claude, Cursor, Windsurf)


Enterprise

Built for production scale

99.9% uptime SLA, guaranteed highest rate limits and task priority for Enterprise accounts.

Exclusive Enterprise-only endpoints, indefinite asset retention (vs. 3 days on self-serve), and volume pricing for committed usage.

Inputs and generated models are never used for training a distinction worth checking against any rival before committing IP to a pipeline.

ISO 27001 certified SOC 2 Type II certified GDPR compliant
Meshy Enterprise API infrastructure overview


Who builds on it

Four core use cases

Game studio using the Meshy API for asset generation

Game studios

Bulk asset generation, rigging and animation via CI/CD pipelines.

3D printing platform using the Meshy API for print-ready models

3D printing platforms

Automated printability checks and repair via the print suite endpoints.

XR application using the Meshy API for 3D content

XR & app developers

On-demand 3D content generation for AR/VR experiences.

Education platform using the Meshy API

AI agents & education

The same endpoints behind the 3D Agent and Agent Skills for coding assistants.



See the engine

What runs behind every API call

Meshy 7 the model behind every generation call

3D Agent built on this same API



Good fit

  • Game studios automating bulk asset generation, rigging and animation in a pipeline.
  • 3D printing platforms needing automated printability checks and repair at scale.
  • App/XR developers generating 3D content on demand inside their own product.
  • Coding-assistant builders integrating 3D generation via the MCP server or Agent Skills.

Know before you build

  • Pro plan or above required for API key access it's not available on the Free tier.
  • 3-day asset retention on self-serve plans; download or re-host promptly, or upgrade to Enterprise.
  • Pay-before-you-go credits are separate from web app subscription credits.
  • Async by design: every endpoint returns a task ID first budget for polling or webhook handling.


FAQ

Frequently asked questions

How do I get a Meshy API key?

Go to Settings → API on meshy.ai. API keys follow the format msy-<random-string> and authenticate via an Authorization: Bearer header. Requires a Pro plan or above.

How much does the API cost?

Pay-before-you-go, separate from web subscription credits. Text to 3D on Meshy 7 costs 20 credits per generation; full per-endpoint pricing is in the table above and on the API pricing page. Legacy Meshy-5/6 models cost roughly 75% less for draft work.

Can I test the API without spending credits?

Yes a free test-mode dummy key (msy_dummy_api_key_for_test_mode_12345678) returns zero-cost mock responses so you can build your integration before generating real assets.

Who owns assets generated through the API?

Paid Meshy users own their generated assets outright. See commercial use and the full license terms.

Is there an MCP server for AI coding assistants?

Yes the Meshy MCP server lets Claude Code, Cursor, Windsurf and other assistants call these same endpoints directly, alongside the open-source Agent Skills repo. Both are covered in depth on the 3D Agent page.

How long are generated assets retained?

3 days maximum on self-serve plans (Pro through Studio); Enterprise plans get indefinite retention. Download or re-host assets promptly on self-serve tiers.

What happens if I exceed my rate limit?

You'll get a 429 Too Many Requests response. Limits scale by plan tier, from 10 max queued tasks on Pro up to 100 on Ultra and a customizable limit on Enterprise see the rate limit table above.

Which model version does the API use by default?

Meshy 7.1, the current default, with Smart Topology output that's animation-ready and 3D-print-ready from a single generation call. Earlier Meshy-5/6 models remain callable for lower-cost draft iteration see the Meshy 7 benchmarks.

What plugins and engine integrations exist?

Official plugins for Blender, Unity, Unreal Engine, Godot, Maya, 3ds Max, Omniverse and Roblox, plus a CLI tool and the MCP server for AI coding assistants.



Start building with the Meshy API

Free test-mode key included build your integration before spending a single credit.

Get API Key →