FAL Model Selector

Quick decision guide for choosing fal.ai models. Read references/ for full parameters.

Key Concepts

Image-to-Video vs Reference-to-Video

Image-to-Video = "animate this exact image"

• Product photo → animate it rotating
• Portrait → make the person smile/blink

Reference-to-Video = "generate video featuring this subject"

• 3 photos of mascot → video of mascot in new scene
• Product from different angles → consistent commercial

Quick Decision Matrix

Video: Text → Video

Video: Image → Video

Video: Reference → Video (Multiple Images)

Video: Extend/Remix

Image: Generate (Text → Image)

Image: Edit

Image: Specialized Apps

Image: Enhancement

Image: Character/Identity

Video: Tools

Other

Detailed References

• references/video-generation.md - 16+ video models (incl. all Veo/Sora endpoints)
• references/image-tools.md - 20+ image models (incl. Gemini/GPT Image)
• references/other-tools.md - TTS, 3D, upscaling, lipsync

Common Patterns

import fal_client

# Product demo video
result = fal_client.subscribe(
    "fal-ai/kling-video/v2.1/pro/image-to-video",
    arguments={
        "prompt": "Product rotates smoothly, camera zooms in",
        "image_url": "https://example.com/product.jpg",
        "duration": "5"
    }
)

# Remove background
result = fal_client.subscribe(
    "fal-ai/birefnet/v2",
    arguments={
        "image_url": "https://example.com/photo.jpg",
        "model": "General"
    }
)

# Studio product photo
result = fal_client.subscribe(
    "fal-ai/image-apps-v2/product-photography",
    arguments={
        "product_image_url": "https://example.com/product.jpg",
        "aspect_ratio": "1:1"
    }
)

# Virtual try-on
result = fal_client.subscribe(
    "fal-ai/image-apps-v2/virtual-try-on",
    arguments={
        "person_image_url": "https://example.com/person.jpg",
        "clothing_image_url": "https://example.com/shirt.jpg"
    }
)

Author

Canlah AI — Run performance marketing without breaking your brand.

• GitHub: github.com/PHY041
• All Skills: clawhub.ai/PHY041