fal-compatible · model serving

Gear Models

Pixal3D, SAM3, and MoGe-2 on our own H200 GPUs, behind a drop-in fal-compatible API. One key, one import — self-hosted models plus the full fal.ai catalog.

Quick start (humans)

Install the client (pulls fal-client from PyPI):

pip install gear-client --extra-index-url https://models.alayalab.ai/install/simple/

Set your key (the only credential you need):
```
export GEAR_KEY=<your internal key>
```

Use it — the only change from stock fal is the import:

import gear_client as fal_client

handler = fal_client.submit("gear/pixal3d", arguments={"input": {
    "image_url": "https://example.com/chair.png",
}})
result = fal_client.result("gear/pixal3d", handler.request_id)
print(result["mesh"]["url"])        # ready-to-download GLB

For your coding agent

Paste this to your agent:

Follow the instructions at
https://models.alayalab.ai/install/agent-instruction.txt
to use the Gear Models 3D generation platform.

Models

gear/pixal3d

Image → high-fidelity textured 3D mesh (Pixal3D, self-hosted H200).

gear/sam3

Image + prompt → detections JSON and mask PNG (SAM3).

gear/moge2

Image → depth, normals, point maps, and optional geometry (MoGe-2).

fal-ai/*

Any other fal model id auto-routes to official fal.ai — same code path.

…

Results come back as ready-to-download URLs, exactly like fal.