Hitem3d Hi3d V3.0 Multi View To 3d API Documentation

Hitem3d Hi3d V3.0 Multi View To 3d API Documentation

Playground

Hi3D V3.0 Multi-View-to-3D reconstructs a detailed 3D mesh from canonical front, back, left, and right reference views, with optional textures, PBR materials, and multiple export formats for game assets, product visualization, 3D design, and production workflows. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

Hi3D V3.0 Multi-view to 3D reconstructs one 3D asset from canonical views of the same object. Provide a front image and optionally add back, left, and right views to improve geometry accuracy, shape consistency, and texture quality.

Why Choose This?

• Multi-view 3D reconstruction
  Build a 3D model from multiple views of the same object for stronger shape consistency.

• Canonical-view workflow
  Start with a required front view, then add back, left, and right images when available.

• Higher geometry reliability
  Multi-view input helps the model better preserve structure, silhouette, and spatial detail.

• Texture and PBR support
  Generate textured output and optional PBR material maps for more production-ready assets.

• Configurable geometry
  Control target face count for different quality and downstream workflow needs.

• Multiple export formats
  Export generated models as glb, obj, stl, fbx, or usdz.

• Shading control
  Adjust de-shading strength to improve texture and material quality.

Parameters

How to Use

1. Upload the front image — Provide a clear front-view image of the object.
2. Add more views optional — Upload back, left, and right images of the same object when available.
3. Choose resolution — Use 2048quality for faster generation or 2048master for higher-quality output.
4. Configure texture optional — Keep enable_texture enabled when textured output is needed.
5. Configure PBR optional — Keep enable_pbr enabled when material maps are required.
6. Set face count optional — Choose a target face count based on your geometry detail needs.
7. Choose export format — Select glb, obj, stl, fbx, or usdz.
8. Adjust shading optional — Use shading to control de-shading strength.
9. Submit — Generate the 3D model and retrieve the output through the standard WaveSpeed prediction response.

Pricing

Pricing is based on selected resolution, texture generation, and PBR generation.

PBR is only generated and priced when texture generation is enabled.

Best Use Cases

• Multi-view object reconstruction — Build a more accurate 3D model from multiple angles of the same object.
• Product 3D assets — Convert product photos into textured 3D models for previews and visualization.
• E-commerce workflows — Create 3D assets for interactive product display.
• Design and prototyping — Turn reference photos into editable 3D assets for iteration.
• Game and real-time pipelines — Generate geometry with configurable face counts for downstream optimization.
• Production-ready exports — Use GLB, OBJ, STL, FBX, or USDZ depending on your workflow.

Pro Tips

• Use images of the same object with consistent lighting, framing, and scale.
• Canonical input order is front, back, left, then right.
• Provide as many clean views as possible for better shape reconstruction.
• Avoid cluttered backgrounds, reflections, heavy shadows, occlusion, or motion blur.
• Use 2048quality for faster iteration and 2048master for higher-quality final output.
• Enable textures when color and surface appearance matter.
• Enable PBR when the output will be used in rendering, visualization, or game-engine workflows.
• Set face_count lower for lightweight assets and higher for more detailed geometry.

Notes

• front_image is required.
• Use canonical views of the same object for the best results.
• PBR is only generated and billed when texture generation is enabled.
• The safety checker follows the upstream default and is not exposed in the public form.
• The generated model file is returned through the standard WaveSpeed prediction response.

Related Models

• Hi3D V3.0 Image to 3D — Generate a 3D model from a single reference image.
• Hi3D V3.0 Multi-view to 3D — Generate a 3D model from multiple reference views.

Authentication

For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result

set -euo pipefail

export WAVESPEED_API_KEY="your-api-key"

REQUEST_BODY=$(cat <<'JSON'
{
  "front_image": "https://interactive-examples.mdn.mozilla.net/media/cc0-images/painted-hand-298-332.jpg",
  "resolution": "2048quality",
  "enable_texture": true,
  "enable_pbr": true,
  "face_count": 2000000,
  "export_format": "glb",
  "shading": 0.5
}
JSON
)

# 1. Submit the prediction.
SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  -X POST "https://api.wavespeed.ai/api/v3/hitem3d/hi3d-v3.0/multi-view-to-3d" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}")

TASK=$(printf '%s' "${SUBMIT_RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
PREDICTION_ID=$(printf '%s' "${TASK}" | jq -r '.id // empty')
if [ -z "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id
' >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 2. Poll until the prediction finishes.
while true; do
  RESPONSE=$(curl --silent --show-error --fail-with-body \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}")
  RESULT=$(printf '%s' "${RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout|deleted) printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    *) sleep 2 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Response Parameters

Result Request Parameters

Result Response Parameters