curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": [
              {"type": "text", "text": "Convert this image into a 3D model. Generate a three-quarter view 3D render showing the object from a slightly elevated angle, revealing depth and volume. Clean studio background, soft lighting, photorealistic 3D rendering."},
              {"type": "image_url", "image_url": {"url": "https://example.com/product-photo.jpg"}}
            ]
          }
    ],
    "stream": false
  }'

Using Python (OpenAI SDK)

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_EACHLABS_API_KEY",
    base_url="https://eachsense-agent.core.eachlabs.run/v1"
)

response = client.chat.completions.create(
    model="eachsense/beta",
    messages=[{
        "role": "user",
        "content": "Convert this image into a 3D model. Generate a three-quarter view 3D render showing the object from a slightly elevated angle, revealing depth and volume. Clean studio background, soft lighting, photorealistic 3D rendering."
    }],
    # Images are included in the message content array above
)

print(response.choices[0].message.content)

With Multiple Reference Images

Provide multiple angles for more accurate reconstruction:

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": [
              {"type": "text", "text": "Convert this object into a 3D model using these reference angles. The first image shows the front, the second shows the side. Generate a clean 3D render from a new angle (three-quarter view from above) that combines both perspectives. Studio lighting, neutral background."},
              {"type": "image_url", "image_url": {"url": "https://example.com/front-view.jpg"}},
              {"type": "image_url", "image_url": {"url": "https://example.com/side-view.jpg"}}
            ]
          }
    ],
    "stream": false
  }'

Images are sent inside messages using the OpenAI multimodal content format. Maximum 4 images per request.

Streaming

Set "stream": true for real-time SSE responses, or "stream": false for complete result in a single response. Streaming is useful for showing progress in UIs; non-streaming is simpler for scripts and automation.

Use Cases

Use Case	Input	Output
Product 3D Preview	product photo	3D render from new angles
Concept Art to 3D	2D character/object art	3D visualization
Real Object Digitization	phone photo	3D model render
Multi-View Generation	single photo	multiple angle renders
E-commerce 360	product image	rotated 3D views
Game Asset Reference	sketch or photo	3D model for reference

Prompt Engineering Tips

Prompt Structure

"Convert this image into a 3D model." + [output angle] + [render style] + [background] + [lighting] + [quality]

Output Angle Keywords

front view, side profile, three-quarter view,
rear view, top-down, bottom-up, isometric,
turntable view (multiple angles in one image)

Render Quality Keywords

photorealistic 3D render, octane render quality,
Blender Cycles render, studio product lighting,
soft ambient occlusion, subsurface scattering,
clean geometry, smooth shading

Examples

Product Photo to 3D

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": [
              {"type": "text", "text": "Convert this product photo into a 3D model render. Show the object from a three-quarter elevated angle that reveals the top and two sides. Maintain exact colors, materials, and proportions from the original photo. Clean white studio background, product photography lighting, photorealistic 3D render."},
              {"type": "image_url", "image_url": {"url": "https://example.com/headphones.jpg"}}
            ]
          }
    ],
    "stream": false
  }'

Character Art to 3D

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": [
              {"type": "text", "text": "Convert this 2D character illustration into a 3D model render. Show a three-quarter view from the right side with slightly elevated camera. Maintain the character design, colors, and proportions from the original art. Stylized 3D render with soft cel-shading, neutral grey background."},
              {"type": "image_url", "image_url": {"url": "https://example.com/character-concept.png"}}
            ]
          }
    ],
    "stream": false
  }'

Multi-View Turntable

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": [
              {"type": "text", "text": "Convert this object into a 3D model and generate a turntable view showing 4 angles in a single image: front, right side, back, and left side. Arranged in a 2x2 grid. Same scale in each view, neutral grey background, consistent studio lighting, clean 3D render."},
              {"type": "image_url", "image_url": {"url": "https://example.com/sneaker-front.jpg"}}
            ]
          }
    ],
    "stream": false
  }'

Architectural Element

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": [
              {"type": "text", "text": "Convert this photo of a decorative column capital into a 3D model render. Show it from an isometric angle with soft directional lighting to emphasize the carved relief details. Clean background, architectural 3D visualization quality, accurate proportions."},
              {"type": "image_url", "image_url": {"url": "https://example.com/column-capital.jpg"}}
            ]
          }
    ],
    "stream": false
  }'

Batch Workflow: 360-Degree Views

OBJECT_IMAGE="https://example.com/product.jpg"

ANGLES=(
  "front view, straight-on"
  "right side profile"
  "three-quarter view from the right, slightly elevated"
  "rear view"
  "three-quarter view from the left, slightly elevated"
  "left side profile"
)

for ANGLE in "${ANGLES[@]}"; do
  curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
    -H "Content-Type: application/json" \
    -H "X-API-Key: $EACHLABS_API_KEY" \
    -d "{
      \"messages\": [{\"role\": \"user\", \"content\": \"Convert this object into a 3D model render. Show it from $ANGLE. Maintain original colors and materials. Clean white background, studio lighting, photorealistic 3D render.\"}],
      \"image_urls\": [\"$OBJECT_IMAGE\"],
      \"stream\": false
    }"
  echo "---"
done

Best Practices for Input Images

Clean background — white or solid color backgrounds produce better results than cluttered scenes.
Good lighting — evenly lit photos reveal more surface detail for reconstruction.
Single object — isolate the subject. Multiple objects confuse the reconstruction.
Sharp focus — blurry or low-resolution images lose detail in 3D conversion.
No heavy occlusion — the more of the object visible, the better the 3D inference.

Common Pitfalls

AI generates rendered images, not actual 3D mesh files. The output is a new-angle 2D render of the interpreted 3D form.
Occluded parts are inferred — the back of an object seen only from the front is guessed. It may not match reality.
Flat or thin objects (cards, stickers, flat art) do not convert meaningfully to 3D.
Reflective and transparent objects (glass, mirrors) are harder to interpret correctly.
Scale is relative — without reference objects, the model cannot infer absolute size.

image-to-3d

Image to 3D

Quick Start

Using curl