Veo 3.1 First-Last-Frame-to-Video

Generate smooth video transitions between two frames using Google DeepMind's Veo 3.1 model via fal.ai.

Prerequisites

• FAL_KEY environment variable must be set (typically in ~/.zshrc)

API Endpoint

POST https://fal.run/fal-ai/veo3.1/first-last-frame-to-video

Parameters

Required

• prompt (string): Text description of how the video should transition between frames
• first_frame_url (string): URL of the first/starting frame
• last_frame_url (string): URL of the last/ending frame

Optional

Usage

cURL

curl --request POST \
  --url https://fal.run/fal-ai/veo3.1/first-last-frame-to-video \
  --header "Authorization: Key $FAL_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "prompt": "Smooth time-lapse transition from day to night as the sun sets",
    "first_frame_url": "https://example.com/daytime-scene.jpg",
    "last_frame_url": "https://example.com/nighttime-scene.jpg",
    "duration": "8s",
    "resolution": "1080p",
    "generate_audio": true
  }'

Python

import fal_client

result = fal_client.subscribe(
    "fal-ai/veo3.1/first-last-frame-to-video",
    arguments={
        "prompt": "Smooth time-lapse transition from day to night as the sun sets",
        "first_frame_url": "https://example.com/daytime-scene.jpg",
        "last_frame_url": "https://example.com/nighttime-scene.jpg",
        "duration": "8s",
        "resolution": "1080p",
        "generate_audio": True
    }
)

# Access the generated video URL
video_url = result["video"]["url"]
print(f"Generated video: {video_url}")

JavaScript

import { fal } from "@fal-ai/client";

const result = await fal.subscribe("fal-ai/veo3.1/first-last-frame-to-video", {
  input: {
    prompt: "Smooth time-lapse transition from day to night as the sun sets",
    first_frame_url: "https://example.com/daytime-scene.jpg",
    last_frame_url: "https://example.com/nighttime-scene.jpg",
    duration: "8s",
    resolution: "1080p",
    generate_audio: true
  }
});

console.log("Generated video:", result.video.url);

Response Format

{
  "video": {
    "url": "https://storage.googleapis.com/.../output.mp4"
  }
}

Examples

1. Time-lapse transition:

   • First frame: Morning cityscape
   • Last frame: Same cityscape at night
   • Prompt: "Time-lapse of the city from morning to night, lights gradually turning on"

2. Seasonal change:

   • First frame: Summer landscape
   • Last frame: Winter landscape (same location)
   • Prompt: "Seasons change from summer to winter, leaves fall and snow covers the ground"

3. Transformation:

   • First frame: Empty room
   • Last frame: Furnished room
   • Prompt: "Furniture magically appears and arranges itself in the room"

4. Growth animation:

   • First frame: Seed in soil
   • Last frame: Full-grown plant
   • Prompt: "Plant grows from seed, stem emerges and leaves unfold"

5. Scene morphing:

   • First frame: Photo of a person
   • Last frame: Artistic painting of the same person
   • Prompt: "Realistic photo gradually transforms into an impressionist painting"

Tips

• Ensure both frames have similar composition for smooth transitions
• Describe the transition motion/action in the prompt
• Works best when both frames share similar subjects or scenes
• Use longer duration (8s) for complex transitions
• The AI will interpolate the motion between the two frames