Reference-to-Video AI Generator
Prompts describe; references show. Reference-to-video lets you attach the things that must stay true, a person's face, a product's shape, a colour palette, a motion from another clip, and then describe everything else. The composer below accepts images, a reference video and a saved character together, and routes to models built for multi-reference generation.
Sign in to see your credit estimate. Creating needs a plan, from $4.99 a week.
Kinds of references and what they control
Style reference
An image whose palette, grain and lighting the model should borrow. Pair it with a prompt about the content.
Character reference
Photos of a person, pet or mascot. Save them once as a character and reuse them in every scene.
Product or object reference
Clean shots of the item from two or three angles keep logos, proportions and materials stable.
Source video
A clip whose motion or framing you want to keep while changing the look.
Multiple references
Combine a character, a product and a style. The router picks models with enough reference slots.
Workflow
- 01
Attach references
Use the image, video and character slots in the composer. Up to several images depending on the routed model.
- 02
Describe the scene
Say what happens and where. Refer to your character by name; the model has its likeness already.
- 03
Pick Character or Cinematic
Character preset prioritises identity consistency; Cinematic prioritises look. Both accept references.
- 04
Generate and iterate
Keep the references, change the prompt, and produce a series that stays consistent.
Getting consistent results
- Three to six references beat one: front, three-quarter and profile for faces; several angles for objects.
- Even lighting in references gives the model less to guess.
- Name the subject in the prompt exactly as it appears in the character.
- Keep clothing and props described the same way across prompts, or bake them into the character description.
Models with reference support
Live from our model registry. Availability and specs update as models change.
Seedance 2.0
Balanced cinematic video with strong motion and reference support.
Gemini Omni
Character-consistent video from reusable references.
Kling Omni
Video-to-video transformation and style transfer.
How model routing works
You never have to read a model comparison to make a clip. Each request is scored against every enabled model in our registry using the capability it needs (text-to-video, image-to-video, first-and-last frame, reference video, character reference), the preset you picked, the aspect ratio and duration, and the model's live health. The best match is submitted; the next best are kept as fallbacks.
Costs are computed the same way. The quote you see before generating is the price for the model that will actually run, including any resolution or audio surcharge, and it is reserved rather than charged until the clip completes.
Private by default, publishable in one tap
Private library
Every creation is stored privately with its prompt, settings and model, so you can rerun, tweak or download it later.
Visibility controls
Publish as public, followers-only, unlisted or private, and decide whether others may remix.
Downloads
Full-resolution files, permanently hosted on our storage, downloadable from the library at any time.
Remix lineage
If you do publish, remixes credit you and appear under your original.
Frequently asked
- How many references can I attach?
- It depends on the routed model; the composer shows the limit for the current selection. Saved characters carry their own reference set, so they only take one slot.
- What is the difference between a reference image and a first frame?
- A first frame becomes the literal opening image of the clip. A reference image informs identity or style without appearing verbatim.
- Can I reference a real person?
- Only with their consent. Creating characters from real people requires an explicit consent confirmation and is covered by the community guidelines.
- Do I have to pick an AI model?
- No. Choose a style preset such as Fast, Cinematic or Story and the router selects the best available model for that intent, aspect ratio, duration and the inputs you attached. If a model is degraded, the router falls back to a comparable one without charging you twice. Advanced users can still pin a specific model.
- How much does a generation cost?
- Every generation is quoted in credits before you confirm, based on the routed model, duration and resolution. Credits are reserved when you start and released automatically if the generation fails, so you only pay for what completes. Plans start at $4.99 a week and include a credit allowance; subscribers can top up with credit packs.
- Are my generations private?
- Yes by default. Everything you generate lands in your private library. You decide per creation whether to publish it to the feed, share it unlisted, restrict it to followers or keep it private, and whether others may remix it.
- Can I download the videos?
- Yes. Every completed generation can be downloaded in full resolution from your library. Files are stored permanently on our own media storage, not on the provider's temporary links.
Related
Ready when you are.
Plans from $4.99 a week. Publish when you're proud of it.