Image to model (H3)

This page lists all supported parameters for image_to_model in the H3 product line. H3 supports two model versions: v3.1-20260211 and v3.0-20250812.

When to use H3

• Choose H3 when you need high-fidelity geometry and broad parameter control.
• If you already have four directional views, use multiview_to_model instead of image_to_model.

Endpoint

Parameters

Required parameters

type: This field must be set to image_to_model.

file: Specifies the image input.

• type: Indicates the file type. Although currently not validated, specifying the correct file type is strongly advised.
• file_token: The identifier you get from upload, please refer to part of Upload directly. Mutually exclusive with url and object.
• url: A direct URL to the image. Supports JPEG and PNG formats with a maximum size of 20MB. Mutually exclusive with file_token and object.
• object (Strongly Recommended): The information you get from upload, please refer to Upload in STS. Mutually exclusive with url and file_token.
  • bucket: Normally it always will be tripo-data.
  • key: The resource_uri from returns.

model_version: v3.1-20260211 or v3.0-20250812.

Common parameters

model_seed: An integer value chosed randomly if not set. This is the random seed for model generation. For model_version>=v2.0-20240919 or Turbo-v1.0-20250506, the seed controls the geometry generation process, ensuring identical models when the same seed is used.

enable_image_autofix: A bool value to determine if optimize the input image to get better generation result. It will take longer time when request. The default value is false.

texture: A bool value to enable texturing, set false to get a base model without any textures. The default value is true. This will cost 10 less credits when set to false, this will not work for version v1.4-20240625.

Advanced parameters

texture_seed: An integer value chosed randomly if not set. This is the random seed for texture generation for model_version>=v2.0-20240919 or Turbo-v1.0-20250506. Using the same seed will produce identical textures. If you want a model with different textures, please use same model_seed and different texture_seed.

texture_alignment: A text value to determine the prioritization of texture alignment in the 3D model. The default value is original_image. All available values are:

• original_image: Prioritizes visual fidelity to the source image. This option produces textures that closely resemble the original image but may result in minor 3D inconsistencies.
• geometry: Prioritizes 3D structural accuracy. This option ensures better alignment with the model’s geometry but may cause slight deviations from the original image appearance.

texture_quality: A text value to control the texture quality. Can be set to detailed or standard. detailed provides high-resolution textures, resulting in more refined and realistic representation of intricate parts. The default value is standard. This will cost 10 additional credits when set to detailed.

Combination rules when texture_quality = detailed:

• texture=false, pbr=false: Upscales the texture to 4K. PBR material will be removed. Only available for v3.0-20250812.
• texture=false, pbr=true: Generates PBR with the current texture.
• texture=true, pbr=false: Regenerates HD texture without PBR.
• texture=true, pbr=true: Regenerates HD texture with PBR.

WARNING

Setting texture_quality=standard with both texture=false and pbr=false is not allowed.

pbr: A bool value to enable pbr, set false to get a model without pbr. The default value is true.

TIP

If this option is set to true, value of texture will be ignored and set to true.

smart_low_poly: A bool value to control if generate low-poly meshes with hand‑crafted topology. Inputs with less complexity work best. The default value is false. This will cost 10 additional credits when set to true.

Caution

There is a possibility of failure for complex models.

quad: A bool value, set true to enable quad mesh output. If quad=true and face_limit is not set, face_limit will default to 10000. Recommended not to exceed 150,000. This will cost 5 additional credits when set to true.

TIP

Enabling this option will force the output to be an FBX model.

face_limit: An integer value which limits the number of faces on the output model. If this option is not set, the face limit will be adaptively determined.

Triangle face limits (default output):

Model Version	Standard (geometry_quality=standard)	Ultra (geometry_quality=detailed)
v3.1-20260211	up to 1,500,000	up to 2,000,000
v3.0-20250812	up to 1,000,000	up to 1,500,000

Quad face limit (quad=true): recommended not to exceed 150,000.

If smart_low_poly=true, it should be 1000~20000; if quad=true as well, it should be 500~10000.

auto_size: A bool value to automatically scale the model to real-world dimensions with the unit in meters. The default value is false.

TIP

This can only be used in textured model.

orientation: A text value. When set orientation=align_image, it will automatically rotate the model to align the original image. The default value is default. This parameter only takes effect when texture=true.

compress: A text value to specify the compression type to apply to the texture. Set to geometry to apply geometry-based compression to optimize the output. By Default we use meshopt compression.

generate_parts: Generate segmented 3D models and make each part editable. The default value is false. This will cost 20 additional credits when set to true.

TIP

Generate_parts is not compatible with texture=true or pbr=true, if you want to generate parts, please set texture=false and pbr=false;

Generate_parts is also not compatible with quad=true, if you want to generate parts, please set quad=false.

export_uv: A bool value to control whether UV unwrapping is performed during generation. The default value is true.

TIP

When set to false, generation speed is significantly improved and model size is reduced. UV unwrapping will be handled during the texturing stage.

geometry_quality: A text value can be set in following option, this will cost 20 additional credits when set to detailed:

• Ultra Mode: Maximum detail for the most intricate and realistic models when setting to detailed.
• Standard Mode: Balanced detail and speed. The default value is standard

Returns

task_id: The identifier for the successfully submitted task.