NEW3DAPIFLUX April 24, 2026 · 5 min read

Turn Any Photo Into a 3D Model — $0.010 Each

We just shipped text-to-3D and image-to-3D on PixelAPI. Upload a photo or write a sentence — get a textured GLB file in 75 seconds. No Blender. No subscriptions. Just an API call.

"We needed 3D product spins for 200 items. Did it in an afternoon with the PixelAPI 3D API. Cost us $2 total."

What It Does

The 3D Generator takes either a text prompt or an image and produces a fully textured 3D model in GLB format — the industry-standard file type that works with Blender, Unity, Unreal Engine, Three.js, and any web browser.

How It Works

Two-step pipeline under the hood:

1. FLUX.1-schnell (10 steps, ~7 seconds) — Generates a clean, well-lit 1024px image from your prompt. Good lighting = better mesh extraction.
2. TripoSR (stabilityai/TripoSR) — Takes the image, estimates depth, runs TSMR (transformer-based mesh reconstruction), and outputs a GLB with embedded color texture.

Total time: 65-90 seconds. VRAM usage peaks at ~12GB (FLUX needs ~7GB, TripoSR needs ~2GB).

Pricing

We built this to be accessible. Here's what you get:

For reference: Meshy.ai charges $0.05-0.15 per model. Our 128-res model is 5-15x cheaper and runs on our own GPU infrastructure — no third-party markups.

API Reference

Text-to-3D

curl -X POST https://api.pixelapi.dev/v1/3d/text-generate \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "prompt=a ceramic blue vase on wooden table, studio lighting" \
  -F "resolution=128"

# Response:
# {"job_id":"abc123","status":"queued","estimated_seconds":75,"credits_used":15}

# Poll for completion:
# GET https://api.pixelapi.dev/v1/3d/text-status/{job_id}
# {"job_id":"abc123","status":"completed","output_url":"https://api.pixelapi.dev/outputs/files/text3d/...model.glb"}

Image-to-3D

curl -X POST https://api.pixelapi.dev/v1/3d/image-generate \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "image=@product_photo.png" \
  -F "resolution=128"

Output Format

Every generation returns two URLs:

• input_image_url — The image AI generated or the one you uploaded
• output_url — The GLB file (downloadable, no auth required)

The GLB uses glTF 2.0 binary format with embedded color texture. No external texture files to manage — the color map is baked into the mesh itself.

Use Cases We've Seen

• E-commerce product showcases — Generate 3D spins for product pages. Upload the photo, get the model, put it in a 360° viewer.
• Game asset prototyping — Describe a prop, get a GLB, drop it in Unity. Iterate fast without a 3D artist.
• Social content — Turn viral products into 3D models for Reels/TikTok. "Show it from every angle" content.
• Architectural viz — Upload a sketch of a furniture piece, get a textured model for presentations.
• AR/VR mockups — Quick 3D models for webAR or Unity-based VR experiences.

Supported Tools

The output GLB opens natively in:

• Blender — Full edit, retexture, retopologize
• Unity — Drag-and-drop into any scene
• Unreal Engine — Import with FBX if needed (use Blender to convert)
• Three.js / WebGL — Load directly in browser, no conversion needed
• Sketchfab — Upload and publish, or use their AR viewer

Limits & Things to Know

• Single image input only — No multi-view reconstruction yet. The underside and back are estimated, not captured.
• Non-photographic input — Cartoons, illustrations, and heavily stylized art may produce lower-quality geometry.
• Not for 3D printing — Mesh topology is not clean enough for direct printing. Use Meshmixer or Blender's decimator first.
• Complex scenes — Multiple objects, busy backgrounds, or low-lit images produce noisy meshes. Use simple, well-lit input.

Get Started

Sign up at pixelapi.dev, get your API key, and make your first call. Free tier includes 100 credits — enough for ~7 text-to-3D generations.

Try it: Live demo · Dashboard · API docs