Atlas Cloud は Dreamina Seedance 2.5 および Seedream 5.0 Pro の米国認定リセラーになりました。
ホーム
探索
Black Forest Labs
Flux.2 Image Models
black-forest-labs/flux-3/extend-video
Atlas Cloud AI 制作スタジオ監督としての可能性を解き放とう。制作をはじめる
BLACKFORESTLABS FLUX 3 Extend Video
動画から動画

BLACKFORESTLABS FLUX 3 Extend Video API by Black Forest Labs

black-forest-labs/flux-3/extend-video
Extend-video

Extend a video, continuing from the source clip final frames, using FLUX 3.

BLACKFORESTLABS FLUX 3 Extend Video は Black Forest Labs が開発したモデルです。Atlas Cloud(運営:Atlas Cloud AI LLC)は本モデルへのアクセスを提供するものであり、その所有者ではありません。すべての商標は各所有者に帰属します。

FLUX 3 Extend Video

FLUX 3 Extend Video continues an existing video clip from where it left off. Provide a source video and a prompt describing how the action continues, and the model generates a seamless new segment that flows naturally from the final frames — with optional native audio.

Why Choose This?

  • Seamless continuation The model reads the source clip's final frames and continues the motion naturally.

  • Native audio generation Produce the extension and its synchronized audio in a single request.

  • Explicit duration required Duration must be set explicitly to validate keyframe positions.

  • Flexible aspect ratios auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, or 9:16.

  • High resolution output Generate the extension in 720p or 1080p.

Parameters

ParameterRequiredDescription
video_urlYesURL of the input video to extend. MP4, under 50 MB and under 15 seconds
promptNoText description of how the video continues from the source clip's final frames
aspect_ratioNoAspect ratio of the generated extension. Default: auto. Options: auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, 9:16
resolutionNoVideo resolution. Default: 720p. Options: 720p, 1080p
durationNoDuration in seconds. Default: 5. Options: 5–20. Cannot be auto — required to validate keyframe positions
generate_audioNoWhether to generate audio. Default: true
safety_toleranceNoSafety level 0 (strictest) to 4 (most permissive). Default: 2
seedNoRandom seed for reproducibility

How to Use

  1. Provide your source video — must be MP4, under 50 MB, and under 15 seconds.
  2. Write your prompt — describe what happens next, continuing from where the clip ends.
  3. Choose resolution — match the source resolution for a consistent result.
  4. Set duration — use auto or specify how many additional seconds to generate.
  5. Run — submit and download your extended video.

Pricing

Billed per second of generated video, by resolution:

ResolutionCost per second
1080p$0.583
720p$0.451

Best Use Cases

  • Long-form Content — Chain multiple extensions to build longer sequences.
  • B-roll Extension — Stretch a short clip to fill a longer edit slot.
  • Iterative Storytelling — Keep a narrative going beyond a single generation.
  • Loop Building — Extend a clip and then join the end back to the start.

Pro Tips

  • Describe the continuation as if you're directing the next shot: "the camera follows the subject as they turn left".
  • Use aspect_ratio: "auto" to preserve the source video's proportions.
  • Keep the prompt consistent in style and tone with the source clip for the smoothest result.
  • Match the resolution to the source video to avoid quality mismatches.

Notes

  • video_url is the only required field; prompt is optional but improves directional control.
  • Source video constraints: MP4 format, under 50 MB, under 15 seconds.
  • duration controls the extension length, not the total output duration.
  • Generation is asynchronous — submit, then poll for the finished video.

類似モデルを探索

ひとつのAPIで、あらゆるメディアAIを。

すべてのモデルを探索