mcpField
F
fal-3d

fal-3d

github.com/fal-ai-communityCreative & media258110 installsUpdated 24w ago
Repository

About fal-3d

fal-3d is a command-line skill for generating 3D models in GLB, OBJ, or PLY formats from text descriptions or images using fal.ai. Model endpoints can be selected, and the README describes finding generation models through search.

Key features

  1. 01Generate 3D models from text
  2. 02Convert images to 3D models
  3. 03GLB, OBJ, and PLY output formats
  4. 04Select model endpoints
  5. 05Pass extra model parameters

Use cases

  1. 01Create a 3D object from a description
  2. 02Build a 3D model from an object image
  3. 03Find models for text-to-3D or image-to-3D

MCP support

The README mentions the `search_models` MCP tool for discovering 3D generation models.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/fal-ai-community/skills --skill fal-3d

After installing, type/fal-3din Claude Code to confirm it loaded.

Bundled filesscripts/generate-3d.sh

Details

From the project README

fal-3d

Generate 3D models (GLB/OBJ/PLY) from text descriptions or images using fal.ai.

Scripts

ScriptPurpose
generate-3d.shGenerate a 3D model from text or image

Usage

Image to 3D
./scripts/generate-3d.sh --image-url "https://example.com/object.jpg" --model fal-ai/hunyuan3d-v3/image-to-3d
Text to 3D
./scripts/generate-3d.sh --prompt "A medieval sword with ornate handle" --model fal-ai/meshy/v6/text-to-3d

Arguments

ArgumentDescriptionRequired
--image-urlURL of image to convert to 3DYes (or --prompt)
--prompt / -pText description for text-to-3DYes (or --image-url)
--model / -mModel endpointNo (default: fal-ai/hunyuan3d-v3/image-to-3d)
--paramExtra param as key=value (repeatable)No

Finding Models

To discover the best and latest 3D generation models, use the search API:

# Search for image-to-3D models
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --category "image-to-3d"

# Search for text-to-3D models
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "text to 3d"

Or use the search_models MCP tool with relevant keywords like "3d", "mesh", "image-to-3d".

Tips

  • For best results from images: use a clear photo with a single object on a plain background
  • Remove background first if needed (use fal-image-edit or fal-generate with birefnet)
  • Simple, well-defined objects work best — complex scenes don't reconstruct well yet
  • 3D generation takes 1-5 minutes — jobs use the queue API

Output Format

{
  "mesh": {
    "url": "https://fal.media/files/.../model.glb",
    "content_type": "model/gltf-binary",
    "file_name": "model.glb"
  }
}

Comments

Markdown supported