vendor_connectors.meshy.tools

`vendor_connectors.meshy.tools`

AI framework tools for Meshy AI 3D generation.

This module provides tools for Meshy AI operations that work with multiple AI agent frameworks. The core functions are framework-agnostic Python functions, with native wrappers for each supported framework.

Module Contents

Classes

`Text3dGenerateSchema`	Pydantic schema for the text3d_generate tool.
`Image3dGenerateSchema`	Pydantic schema for the image3d_generate tool.
`RigModelSchema`	Pydantic schema for the rig_model tool.
`ApplyAnimationSchema`	Pydantic schema for the apply_animation tool.
`RetextureModelSchema`	Pydantic schema for the retexture_model tool.
`ListAnimationsSchema`	Pydantic schema for the list_animations tool.
`CheckTaskStatusSchema`	Pydantic schema for the check_task_status tool.
`GetAnimationSchema`	Pydantic schema for the get_animation tool.

Functions

`_extract_result_fields`	Extract common fields from Meshy API result objects.
`text3d_generate`	Generate a 3D model from text description.
`image3d_generate`	Generate a 3D model from an image.
`rig_model`	Add skeleton/rig to a static 3D model.
`apply_animation`	Apply animation to a rigged model.
`retexture_model`	Apply new textures to an existing model.
`list_animations`	List available animations from the Meshy catalog.
`check_task_status`	Check status of a Meshy task.
`get_animation`	Get details of a specific animation.
`get_langchain_tools`	Get all Meshy tools as LangChain StructuredTools.
`get_crewai_tools`	Get all Meshy tools as CrewAI tools.
`get_strands_tools`	Get all Meshy tools as plain Python functions for AWS Strands.
`get_tools`	Get Meshy tools for the specified or auto-detected framework.

Data

`TOOL_DEFINITIONS`

API

class vendor_connectors.meshy.tools.Text3dGenerateSchema

Bases: pydantic.BaseModel

Pydantic schema for the text3d_generate tool.

prompt : str

‘Field(…)’

art_style : str

‘Field(…)’

negative_prompt : str

‘Field(…)’

target_polycount : int

‘Field(…)’

enable_pbr : bool

‘Field(…)’

class vendor_connectors.meshy.tools.Image3dGenerateSchema

Bases: pydantic.BaseModel

Pydantic schema for the image3d_generate tool.

image_url : str

‘Field(…)’

topology : str

‘Field(…)’

target_polycount : int

‘Field(…)’

enable_pbr : bool

‘Field(…)’

class vendor_connectors.meshy.tools.RigModelSchema

Bases: pydantic.BaseModel

Pydantic schema for the rig_model tool.

model_id : str

‘Field(…)’

wait : bool

‘Field(…)’

class vendor_connectors.meshy.tools.ApplyAnimationSchema

Bases: pydantic.BaseModel

Pydantic schema for the apply_animation tool.

model_id : str

‘Field(…)’

animation_id : int

‘Field(…)’

wait : bool

‘Field(…)’

class vendor_connectors.meshy.tools.RetextureModelSchema

Bases: pydantic.BaseModel

Pydantic schema for the retexture_model tool.

model_id : str

‘Field(…)’

texture_prompt : str

‘Field(…)’

enable_pbr : bool

‘Field(…)’

wait : bool

‘Field(…)’

class vendor_connectors.meshy.tools.ListAnimationsSchema

Bases: pydantic.BaseModel

Pydantic schema for the list_animations tool.

category : str

‘Field(…)’

limit : int

‘Field(…)’

class vendor_connectors.meshy.tools.CheckTaskStatusSchema

Bases: pydantic.BaseModel

Pydantic schema for the check_task_status tool.

task_id : str

‘Field(…)’

task_type : str

‘Field(…)’

class vendor_connectors.meshy.tools.GetAnimationSchema

Bases: pydantic.BaseModel

Pydantic schema for the get_animation tool.

animation_id : int

‘Field(…)’

vendor_connectors.meshy.tools._extract_result_fields(result: object) → dict[str, object]

Extract common fields from Meshy API result objects.

Safely extracts status, model_url, and thumbnail_url from result objects, handling missing attributes gracefully.

Args: result: A Meshy API result object (Text3DResult, Image3DResult, etc.)

Returns: Dict with status, model_url, and thumbnail_url fields

vendor_connectors.meshy.tools.text3d_generate(prompt: str, art_style: str = ‘realistic’, negative_prompt: str = ”, target_polycount: int = 30000, enable_pbr: bool = True) → dict[str, Any]

Generate a 3D model from text description.

Args: prompt: Detailed text description of the 3D model (max 600 chars) art_style: One of: realistic, sculpture. For ‘sculpture’, set enable_pbr=False as sculpture style generates its own PBR maps. negative_prompt: Things to avoid in the generation target_polycount: Target polygon count enable_pbr: Enable PBR materials. API defaults to False; we default to True for better realistic renders. Set False for sculpture.

Returns: Dict with task_id, status, model_url, and thumbnail_url

vendor_connectors.meshy.tools.image3d_generate(image_url: str, topology: str = ”, target_polycount: int = 15000, enable_pbr: bool = True) → dict[str, Any]

Generate a 3D model from an image.

Args: image_url: URL to the source image topology: Mesh topology (“quad” or “triangle”), empty for default target_polycount: Target polygon count enable_pbr: Enable PBR materials

Returns: Dict with task_id, status, model_url, and thumbnail_url

vendor_connectors.meshy.tools.rig_model(model_id: str, wait: bool = True) → dict[str, Any]

Add skeleton/rig to a static 3D model.

Args: model_id: Task ID of the static model to rig wait: Whether to wait for completion (default True)

Returns: Dict with task_id and status

vendor_connectors.meshy.tools.apply_animation(model_id: str, animation_id: int, wait: bool = True) → dict[str, Any]

Apply animation to a rigged model.

Args: model_id: Task ID of the rigged model animation_id: Animation ID from the Meshy catalog (integer) wait: Whether to wait for completion (default True)

Returns: Dict with task_id, status, and glb_url

vendor_connectors.meshy.tools.retexture_model(model_id: str, texture_prompt: str, enable_pbr: bool = True, wait: bool = True) → dict[str, Any]

Apply new textures to an existing model.

Args: model_id: Task ID of the model to retexture texture_prompt: Description of the new texture/appearance enable_pbr: Enable PBR materials wait: Whether to wait for completion (default True)

Returns: Dict with task_id, status, and model_url

vendor_connectors.meshy.tools.list_animations(category: str = ”, limit: int = 50) → dict[str, Any]

List available animations from the Meshy catalog.

Args: category: Optional category filter (Fighting, WalkAndRun, etc.) limit: Maximum number of results

Returns: Dict with count, total, and list of animations

vendor_connectors.meshy.tools.check_task_status(task_id: str, task_type: str = ‘text-to-3d’) → dict[str, Any]

Check status of a Meshy task.

Args: task_id: The Meshy task ID task_type: Task type (text-to-3d, rigging, animation, retexture)

Returns: Dict with status, progress, and model_url if complete

vendor_connectors.meshy.tools.get_animation(animation_id: int) → dict[str, Any]

Get details of a specific animation.

Args: animation_id: The animation ID number

Returns: Dict with animation details

vendor_connectors.meshy.tools.TOOL_DEFINITIONS

None

vendor_connectors.meshy.tools.get_langchain_tools() → list[Any]

Get all Meshy tools as LangChain StructuredTools.

Returns: List of LangChain StructuredTool objects for Meshy operations.

Raises: ImportError: If langchain-core is not installed.

vendor_connectors.meshy.tools.get_crewai_tools() → list[Any]

Get all Meshy tools as CrewAI tools.

Returns: List of CrewAI BaseTool objects for Meshy operations.

Raises: ImportError: If crewai is not installed.

vendor_connectors.meshy.tools.get_strands_tools() → list[Any]

Get all Meshy tools as plain Python functions for AWS Strands.

Strands agents work with plain Python functions that have type hints. The functions are returned as-is since they already have proper signatures.

Returns: List of callable functions for Strands agents.

vendor_connectors.meshy.tools.get_tools(framework: str = ‘auto’) → list[Any]

Get Meshy tools for the specified or auto-detected framework.

This is the recommended entry point - it auto-detects which AI framework is installed and returns tools in the appropriate format.

Args: framework: Framework to use. Options:

“auto” (default): Auto-detect based on installed packages
“langchain”: Force LangChain StructuredTools
“crewai”: Force CrewAI tools
“strands”: Force plain functions for Strands
“functions”: Force plain functions (alias for strands)

Returns: List of tools in the appropriate format for the framework.

Raises: ImportError: If the requested framework is not installed. ValueError: If an unknown framework is specified.

Example:

Auto-detect (recommended)

tools = get_tools()

# Force specific framework
tools = get_tools("langchain")
tools = get_tools("crewai")