floyo logo
Pricing
floyo logo
Pricing

Pixverse C1 Reference to Video

Upload up to 7 reference images, tag each one, and generate a video that composes your characters, objects, and backgrounds into one scene with Pixverse C1.

28

Generates in about -- secs

Nodes & Models

LoadImage
CreateVideo
SaveVideo
VideoToFrames

Generate a video scene from multiple reference images. Upload up to 7 pictures (a character, an object, a setting) and tag each one. Write a prompt that calls them by name, and Pixverse C1 composes them into a single video with their identities preserved.

Clips run 1 to 15 seconds at up to 1080p, with optional audio.

How do you use multi-reference video generation with Pixverse C1?

Upload 1 to 7 reference images, give each one a short name (like hero, cafe, product), and tag each as a subject or background. In your prompt, call them using @name. Pixverse C1 builds a video scene that keeps every reference identifiable and consistent throughout the clip.

Reference Images (1 to 7) Drop in clean, well-lit shots with a single clear subject per image. More references mean more for the model to track, so 2 to 3 usually gives the cleanest compositions. Past that, attention spreads thin and identities start to blur.

Ref Names Descriptive beats generic. "@hero" and "@cafe" read cleaner in prompts than "@ref1" and "@ref2". Whatever you name them, you have to use the exact same name in your prompt for the model to pick them up.

Type (subject or background) Subject for characters, products, or objects you want to see in the scene. Background for environments, locations, or scene images the action should happen in. Tag them right and the model knows what to composite where. Mis-tagging a location as a subject is a common reason scenes come out wrong.

Prompt Write the scene. Call your references by name with an @ prefix. Example: "@alex and @maya order coffee at @cafe while rain taps the window." Motion is still driven by text. The references lock identity, not action.

Resolution Want to test a scene composition? Use 360p or 540p for quick iterations. Want final delivery? Go 720p or 1080p. The tradeoff is render speed and cost against detail.

Duration Anywhere from 1 to 15 seconds. Short clips (3 to 5 seconds) render faster and are easier to iterate on. Longer clips give the scene room to breathe but cost more compute.

Audio Switch Off by default. Flip it on when the scene has ambient sound worth generating: crowds, outdoor settings, action sequences. Leave it off for silent loops or anything where you plan to add dialogue and music in post.

Aspect Ratio 16:9 for widescreen and YouTube. 9:16 for phones and Reels. 1:1 for social posts. Match your delivery platform.

What is Pixverse C1 Reference to Video good for?

Shots where specific characters, products, or locations have to appear and stay recognizable. Good for brand characters in new scenes, a cast meeting in a chosen location, or product placement in a custom environment. Anywhere a single start-image workflow can't handle multiple identities in one frame.

Use it for brand characters that need to stay on-model across a series of clips, pre-visualization where you want to test an actor in a location you haven't shot yet, or multi-character scenes where two or three specific faces have to interact in the same frame. Product teams use it to drop a real product into a generated environment without any 3D work.

If you only have one image to animate, Pixverse C1 Image-to-Video WaveSpeedAI is the simpler tool. If you don't need identity locked at all, the text-to-video version is faster and cheaper.

FAQ

How many reference images can Pixverse C1 use at once? Up to 7 per generation. 2 to 3 usually produces the cleanest results because the model splits attention across references. Past that, individual identities start to blur together. Save the extra slots for runs where multiple locked subjects are non-negotiable.

How long can videos be with Pixverse C1 Reference to Video? 1 to 15 seconds per generation. Short clips (3 to 5 seconds) render quickly and work well for testing a scene composition before committing to a longer render at higher resolution.

Should I tag references as subject or background in Pixverse C1? Characters, objects, and products go under subject. Locations, environments, and scene images go under background. The tag tells Pixverse C1 where to place each reference in the composite. Mis-tagging a background as a subject is a common reason generations look off.

Does Pixverse C1 generate audio with the video? Only when the audio switch is on. It works best for scenes with environmental sound like outdoor settings, crowds, and action sequences. Leave it off for silent loops, stock b-roll, or anything where dialogue and music are coming in post.

How to run Pixverse C1 Reference to Video online? You can run Pixverse C1 Reference to Video online through Floyo. No installation, no setup. Open the workflow in your browser, upload your inputs, and hit run. Free to try.

Read more

N