# Flux 3 Image to Video - Core.Today AI API

> FLUX.3 is Black Forest Labs' frontier video model. This endpoint animates a single still image into coherent, natural motion at 720p or 1080p for 5-20 seconds, with synchronized audio generated alongside the video and a tunable safety tolerance.

- **Provider**: Black Forest Labs
- **Model ID**: blackforestlabs/flux-3/image-to-video
- **Category**: Video Generation
- **Credits**: 1980 per 5s 720p video — billed per second (720p 395/s, 1080p 674/s)
- **Speed**: Medium
- **Quality**: Ultra

## Features
- 1080p output — the highest resolution tier among our image-to-video endpoints with native audio
- 5-20 second durations, billed per second
- Audio generated alongside the video rather than dubbed afterwards
- Eight aspect ratios from 21:9 to 9:16, or auto
- Adjustable safety tolerance from 0 (strictest) to 4 (most permissive)

## Use Cases
- 1080p hero clips for landing pages and paid placements where 720p is not enough
- Animating key art or a product still into a short cinematic shot
- Ambient b-roll generated from an existing photograph, with matching ambient sound
- Motion tests at 720p before committing to a 1080p final render

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **prompt**: string - The text prompt describing the video you want to generate.
- **image_url**: string - URL of the image the video starts from (PNG, JPEG, or WebP).
- **duration**: integer - Duration of the generated video in seconds (5-20). Required — upstream also accepts "auto", but the resulting length is unknown at request time and cannot be billed accurately, so this gateway requires an explicit duration.

### Optional
- **resolution**: string (default: 720p) - Resolution of the generated video.
  Options: 720p, 1080p
- **aspect_ratio**: string (default: auto) - Aspect ratio of the generated video. 'auto' lets the model choose.
  Options: auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, 9:16
- **generate_audio**: boolean (default: true) - Whether to generate audio for the video.
- **safety_tolerance**: integer (default: 2) - The safety tolerance level for the generated video. 0 is the strictest and 4 is the most permissive.

## Examples

### Single continuous shot from a still
Animate a photograph into one unbroken 5-second take at 720p, with ambient sound generated alongside.
```json
{
  "model": "blackforestlabs/flux-3/image-to-video",
  "input": {
    "prompt": "A red panda walks along a mossy log in a sunlit forest, one continuous unbroken shot. Ambient birdsong and rustling leaves.",
    "image_url": "https://storage.googleapis.com/falserverless/example_inputs/veo31_i2v_input.jpg",
    "duration": 5,
    "resolution": "720p",
    "aspect_ratio": "auto",
    "generate_audio": true
  }
}
```

### 1080p hero clip
Render the final cut at 1080p in a 16:9 frame once the motion has been validated at 720p.
```json
{
  "model": "blackforestlabs/flux-3/image-to-video",
  "input": {
    "prompt": "Slow dolly-in on a ceramic coffee cup as steam rises, morning light through a window, shallow depth of field.",
    "image_url": "https://example.com/product-still.jpg",
    "duration": 8,
    "resolution": "1080p",
    "aspect_ratio": "16:9",
    "generate_audio": true
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Validate the motion at 720p first — 1080p costs roughly 1.7x per second for the same duration.
- Billing is per second of output, so an 8-second clip costs about 60% more than a 5-second one.
- Describe one continuous camera move rather than a sequence of cuts; this endpoint animates a single still and holds up best on unbroken shots.
- safety_tolerance defaults to 2. Lower it for brand-safe placements; raise it only when a legitimate prompt is being over-filtered.

## Documentation
https://fal.ai/models/blackforestlabs/flux-3/image-to-video