Imagera AI - AI content creation platform for generating images, cloning voices, and enhancing videos. Privacy Policy | Terms

Illustration: a model beside pinned photographs of her from every angleIllustration

Imagera IDs

Character ID One reference. Every shot.

No training. Works with every model that takes a picture.

Photos or a few words become one 2K or 4K reference sheet — angles, expressions and details in a single image. Type @ to use it in Sandbox, or add it as an input to any edit, reference or video model.

  • 2K or 4KOne reference sheet per ID
  • 6Photos at most, or a few words
  • 6Kinds of ID: character, object, location, outfit, digital product, property
  • 0Training runs. The sheet is a picture.

Start from one

Examples

IDs are saved to your account and follow you into every tool.

Actors

Ready-made people for your videos and photos.

352 ready-made actors. See all opens the library for every filter and a closer look.

Three steps

How it works

A sheet is an ordinary picture, so every model that takes one can use it.

  1. The one reference photo the Juniper sheet was made from

    Add references

    Up to 6 photos or generations — or describe it in words.

  2. Character reference sheet of Juniper, made with Character ID from one photo

    Get a 2K or 4K sheet

    Angles, expressions and details in one collage, lit and scaled the same.

  3. Juniper walking through a rainy night market — made in Sandbox with @juniper-ref on ChatGPT 2.5 Flare Edit

    Use it anywhere

    Type @handle in Sandbox, pick it in the upload modal, or pass it to Canvas and the API.

Illustration: a man recording his voice at a microphoneIllustration

New: voice

Give them a voice. Hear it in every clip.

Upload a recording, or record up to 13 seconds on the page.

A Character ID can keep one voice sample. Video models that take a voice reference use it as the voice of that character, so they speak your lines in it. An ID without a voice works as before.

Use your own voice, or one you have permission to use.

Every tool

Use an ID

The same sheet works in every tool.

All skills
Juniper walking through a rainy night market — made in Sandbox with @juniper-ref on ChatGPT 2.5 Flare EditMade with @juniper-ref

Sandbox · @handle

Type @ in the prompt of any model that takes a picture. The sheet joins the inputs.

Made from the sheet

Video · Warlord Pro Reference

“Make a video” opens the reference model with the sheet loaded — the same face in every clip.

The Character ID skill, with Open in Canvas, Ask Superagent and SKILL.md for agentsProduct capture

Canvas, Superagent & MCP

Run the ID skills as Canvas apps, load them in Superagent with /, or give them to your own agent.

Good to know

Questions

Is a Character ID a trained model?
No. It is one high-resolution reference sheet. Any model that accepts an input image uses it directly, so it works with new models the day they launch.
Which models can use an ID?
Any image edit, reference or image-to-video model — for example ChatGPT 2.5 Edit, Typeset 3 Edit and Imagera Warlord Pro Reference. Type @ in Sandbox or pick the ID in the upload modal.
What if ChatGPT Image 2 declines my references?
Typeset 3 makes the sheet instead, automatically. Every sheet is measured: one that comes back smaller than the size you picked is upscaled to it.
Is my character mine to keep?
Yes. An ID is an ordinary picture: download the sheet from the ID to keep your own copy, and use it with any model that takes a reference image, on Imagera or anywhere else.
Can I make IDs of objects, places and outfits?
Yes. Object, Location and Outfit IDs use sheets laid out for their subject: product angles and macro details, a place from several angles, garments front, back and flat.
Illustration: the same model walking through a night marketIllustration

Start here

Make the first one. Use it in every shot.

Add references or describe one. The sheet takes about a minute.