Higgsfield Cinematic Studio Video API

Image to Video

Professional cinematic videos with camera movements. AI sound effects and slow motion support.

·Features·FAQ

Documentation

Prompt

Text description to guide the video generation

Shot Mode

single: one continuous shot, multi_shot_auto: AI-controlled cuts, multi_shot_manual: define each shot

Reference Images

Up to 3 reference images. Use <<<image_1>>>, <<<image_2>>>, <<<image_3>>> in prompt.

0/3

Click or drag & dropPNG, JPG, WEBP, GIF · Max 100MB

Start Image

Starting image for the video (required)

0/1

Click or drag & dropPNG, JPG, WEBP, GIF · Max 100MB

End Image

End image for transitions (disables camera effects when set)

0/1

Click or drag & dropPNG, JPG, WEBP, GIF · Max 100MB

Duration

Video duration in seconds

Aspect Ratio

Output video aspect ratio

Resolution

Output video resolution

Genre

Cinematic genre style

Camera Movement ID

Camera movement preset ID (optional)

Sound

Enable AI-generated sound effects (only available without end image)

Seed

Random seed for reproducible results

Output

Your generated video will appear here

Features

What Higgsfield Cinematic Studio Video API offers

Shot modes: single continuous take, multi_shot_auto for AI-driven cuts, or multi_shot_manual when you script each beat

Durations from 3 to 12 seconds (slider in the playground)

720p or 1080p output resolution

Eight aspect ratios including 21:9 cinematic plus square, portrait, and landscape crops

Optional reference image bundle (up to three uploads) referenced via <<<image_1>>> style tokens inside the prompt

Optional start_image_url and end_image_url; enabling an end frame automatically disables sound generation per API rules

Genre presets (auto, action, horror, comedy, western, suspense, intimate, spectacle) plus optional camera_movement_id UUID

Optional AI sound toggle, optional seed (0 to 999999), JSON generate and poll flow

Use cases

Built for

Primary

Trailer beats - Multi-shot modes when you need coverage without manual editing

Product storytelling - Start and end frames for before and after reveals

Vertical social - 9:16 exports with genre styling for platform-native motion

Pitch sizzles - Ultrawide 21:9 boards for theater-style reviews

Real estate flyovers - Reference stills with <<<image_n>>> tokens for room context

Sound-off renders - Omit end frames so the sound flag can stay enabled

FAQ

About Higgsfield Cinematic Studio Video API

You can combine prompt text, shot_mode, optional uploads (up to three reference stills, optional dedicated start and end frames), duration, aspect_ratio, resolution, genre, optional camera_movement_id, optional sound, and optional seed.

When end_image_url is present the API forces sound off because the pipeline prioritizes the transition between frames.

Upload up to three files and mention <<<image_1>>>, <<<image_2>>>, or <<<image_3>>> inside the prompt so the model knows how to use each asset.

Usage records against higgsfield-ai/cinematic-studio-video. Duration, resolution, and sound all influence compute, so check the published rate card for credits per second.