What is reference to video AI?

It is a generation workflow that uses media or links in addition to text, giving the model stronger guidance for subjects, products, scenes, motion, sound, and style.

Which reference types can I use?

MiniMax H3 and Seedance accept images, video, and audio in the synced model configuration. Wan can additionally accept supported links. Happy Horse currently uses image references.

How many references can I add?

Limits differ by model and media type. The page uses the limits synchronized from the Flaq SaaS Template, while the Flaq API remains the final validator.

How should I mention reference assets in the prompt?

Describe them by order and type, such as first image, second clip, or audio reference, and state the specific role each asset should play.

Are local references private?

They are selected locally, but generation requires uploading them to a signed public URL and sending those URLs to Flaq AI. Do not use sensitive media unless this processing model is acceptable.

Does this workflow require API credits?

Yes. The Free Video AI interface needs no site signup, but Flaq AI charges the connected Client Key according to its current model pricing.

Reference to Video AI workflow combining visual and audio inputs

Free Video AI / Multimodal direction

Reference to Video AI — Guide Characters, Style, Motion, and Sound

Combine images, video clips, audio, and supported links to direct a new AI video with Seedance 2.5, MiniMax H3, Wan 3.0, or Happy Horse through the Flaq API.

Multimodal input4 model familiesLocal encrypted keyAutomatic history

Create with References

Direct a Video with References

References into direction

Reference-Guided Video Examples

Reference media can anchor identity, products, environments, movement, voice, rhythm, or visual style when paired with an explicit prompt.

Omni-Reference Generation

Mixed reference types help describe a subject and audiovisual direction beyond what text alone can specify.

First and Last Frame Direction

Visual anchors can establish the opening composition and desired destination while the model creates the transition.

Examples are provided by the open-source Flaq SaaS Template or the MiniMax H3 repository. Results vary by model, prompt, inputs, settings, safety policy, and service availability.

A Multimodal Reference Workspace

Bring reference media and URLs into one prompt-driven form with model-aware limits and a consistent Flaq API task flow.

Images, Video, and Audio

Upload supported media or paste public URLs, then explain exactly how each asset should influence the result.

Model-Specific Inputs

The form hides video, audio, and link controls when the active reference model does not accept them.

Persistent Local Task History

Submitted task IDs and returned URLs stay in this browser so you can revisit recent results without a site account.

Flaq AI model APIs

Reference to Video Use Cases

Reference control is useful when a production must follow an existing person, product, scene, audio cue, or visual language.

Consistent Characters

Use clear portraits or character art to guide identity, wardrobe, and styling across a new scene.

UGC and Product Ads

Reference a product, presenter, or brand asset while directing a new demonstration or campaign variation.

Style and Motion Transfer

Use visual clips to communicate framing, pace, movement, atmosphere, or animation language.

Audio-Led Scenes

Where supported, add voice or sound references and describe how the new video should respond to them.

How to Generate from References

Select the model first, add only compatible references, and make every asset's role explicit in the prompt.

Start Reference to Video

1. Choose the Reference Model

Compare supported images, clips, audio, links, duration, resolution, sound, and reference limits.

2. Add and Explain References

Upload files or paste public URLs, then identify how each reference controls the subject, action, style, camera, or sound.

3. Generate and Inspect

Submit with your Flaq Client Key, follow automatic status updates, and review the output for consistency before production use.

Explore More Video AI Workflows

Move from an idea to an animated frame or a reference-directed production without leaving the Free Video AI toolkit.

Video AI tool

Text to Video

Generate a new clip from a detailed written direction.

Open tool Video AI tool

Image to Video

Animate a still image while guiding movement and camera behavior.

Open tool

Reference to Video AI FAQ

Video Guides