Meshy MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MESHY_API_KEY | Yes | Your Meshy API key from meshy.ai/settings/api |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| text_to_3d_createA | Generate a 3D model from a text prompt. Returns a task ID to poll for results. Use mode 'preview' first, then 'refine' with the preview_task_id. |
| text_to_3d_getA | Check the status of a text-to-3D task. Use wait_for_task to poll until complete. |
| text_to_3d_listB | List text-to-3D tasks with pagination. |
| text_to_3d_deleteC | Delete a text-to-3D task. |
| image_to_3d_createB | Generate a 3D model from a single image. The image_url must be publicly accessible or a base64 data URI. |
| image_to_3d_getA | Check the status of an image-to-3D task. |
| image_to_3d_listC | List image-to-3D tasks. |
| image_to_3d_deleteB | Delete an image-to-3D task. |
| multi_image_to_3d_createA | Generate a 3D model from multiple images (1-4). Provide publicly accessible URLs. |
| multi_image_to_3d_getB | Check the status of a multi-image-to-3D task. |
| multi_image_to_3d_listC | List multi-image-to-3D tasks. |
| multi_image_to_3d_deleteB | Delete a multi-image-to-3D task. |
| remesh_createB | Remesh and export a 3D model into various formats (glb, fbx, obj, usdz, blend, stl). Provide either input_task_id or model_url. |
| remesh_getB | Check the status of a remesh task. |
| remesh_listC | List remesh tasks. |
| remesh_deleteB | Delete a remesh task. |
| retexture_createB | Apply new textures to a 3D model. Provide either input_task_id or model_url, and either text_style_prompt or image_style_url. |
| retexture_getB | Check the status of a retexture task. |
| retexture_listC | List retexture tasks. |
| retexture_deleteC | Delete a retexture task. |
| rigging_createA | Auto-rig a humanoid 3D model for animation. Requires GLB format, max 300k faces (use remesh first if over). Provide either input_task_id or model_url. |
| rigging_getB | Check the status of a rigging task. |
| rigging_deleteC | Delete a rigging task. |
| animation_createA | Apply an animation to a rigged model. Requires a completed rigging task. Common categories: DailyActions (0=Idle), WalkAndRun (1=Walking), Fighting (4=Attack), Dancing (22-24), BodyMovements. See Meshy animation library for full list of 500+ action IDs. |
| animation_getB | Check the status of an animation task. |
| animation_deleteB | Delete an animation task. |
| text_to_image_createB | Generate an image from text (useful as input for image-to-3D). Models: 'nano-banana' or 'nano-banana-pro'. |
| text_to_image_getB | Check the status of a text-to-image task. |
| text_to_image_listB | List text-to-image tasks. |
| text_to_image_deleteB | Delete a text-to-image task. |
| image_to_image_createB | Transform and edit an existing image using reference images and a text prompt. Models: 'nano-banana' or 'nano-banana-pro'. |
| image_to_image_getA | Check the status of an image-to-image task. |
| image_to_image_listC | List image-to-image tasks. |
| image_to_image_deleteB | Delete an image-to-image task. |
| wait_for_taskA | Poll a task until it reaches a terminal state (SUCCEEDED, FAILED, or CANCELED). Returns the final task result with download URLs. Use this instead of manually calling _get in a loop. |
| get_balanceA | Check your Meshy account credit balance. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 36 tools
Each tool targets a distinct resource (e.g., text_to_3d, image_to_3d, retexture) and action (create, get, delete, list), with no overlap. Helper tools like wait_for_task and get_balance are clearly separate. An agent can easily distinguish purposes.
All tools follow a strict 'category_action' snake_case pattern (e.g., text_to_3d_create, image_to_image_list). No mixing of conventions, and the naming is predictable and readable.
36 tools is high, but justified by the breadth of the domain (text, image, multi-image, retexture, remesh, rigging, animation) where each feature requires CRUD operations plus a few utilities. Slightly above the ideal range, but not excessive given the complexity.
The tool set covers the full lifecycle for each 3D generation/modification feature: create, get status, delete, and list. Additionally, there are helpers (wait_for_task, get_balance) and supporting tools (text_to_image, image_to_image). No obvious gaps for the stated purpose of 3D asset creation and editing.