Real-Person Photo Getting Flagged? Try Idoli’s Character Workflow

You upload a portrait, describe a new outfit or a short video, and press generate. Instead of a result, you receive a restriction related to the real person in the photo. You change the wording and try again, but your creative idea still has not reached the screen.

If that is where you are getting stuck, Idoli’s character workflow is worth understanding. Rather than using the original real-person photo directly for every generation, first turn the portrait into a character and save it. Then reference the saved character in image or video mode.

For supported creations, this is Idoli’s route around the real-person-photo restriction associated with directly using the photo. The important step is not simply uploading a different file: it is creating, saving, and then referencing the character. Here is how that opens up your next project.

Why the saved character is the key

In this workflow, a character is more than an image you keep in a library. It becomes a reusable subject for new image and video creations.

Start with a portrait, establish the character’s appearance, and save the result. After that, your next prompt can focus on what you want to create around that character: a different wardrobe, a new location, a cinematic atmosphere, or a short performance.

The sequence matters:

Real-person photo → Generate a character → Fine-tune → Save → Reference the character in image or video mode.

This is the character-based route Idoli provides for real-person-photo projects. It addresses the direct-photo issue without requiring you to reconstruct the person’s appearance from a long text description every time.

One character, more creative directions

The original photo does not have to define every future outfit and background. Once you have saved the character, you can reference it in a new scene and describe a different visual direction.

Imagine starting with a portrait of an adult woman with copper-red hair. You could use the saved character for a poolside fashion image, a rain-lit city portrait, and a garden video. Each project uses the same selected character, while the prompt changes the setting and performance.

That is the practical advantage: the character is the starting point, not the limit of the scene. You spend less of each prompt describing who the person is and more describing what you want to make.

Reference-based generation still deserves a visual check. Review the face and other recognizable features in each result rather than assuming every detail will remain identical.

Create, refine, and save your first character

Open character-generation mode and upload a portrait you have permission to use. A clear, evenly lit single-person photo makes the generated features easier to inspect.

Generate the character, then check the face, hair, and other defining details. If an adjustment is needed, click Fine-tune on the result. For example:

Change the hair to a darker copper-red shade. Keep the facial features and hairstyle unchanged.

Character-generation mode currently does not have conversation memory, so use Fine-tune for these detail adjustments.

When the result looks right, save the character. You are now ready to use it in image and video mode. Do not stop at generating the character: saving it and selecting it in the next mode completes the workflow.

Reference the character for a new image

Switch to image-generation mode and reference the saved character. Describe the photograph you want to create rather than repeating the full original portrait.

For an adult character in a fashion-editorial scene, try:

Create a realistic fashion-editorial photograph of the referenced adult character beside a rooftop swimming pool at sunset. She wears an opaque black one-piece swimsuit and an open ivory linen shirt. Frame her from mid-thigh upward, standing naturally beside a pale stone railing, one hand resting on it. Warm evening light catches her copper-red hair, while the distant city skyline falls softly out of focus. Elegant composition, believable skin texture, and natural proportions.

This gives the character a new outfit, location, pose, and photographic treatment. The creative direction comes from the prompt, while the saved character provides the subject reference.

To explore a different mood, reference the same character and try:

Create a cinematic waist-up portrait of the referenced adult character on a rain-wet city street at night. She wears a dark green tailored coat over a black top. Her body faces slightly away while her head turns toward the viewer with a calm, attentive expression. Pink signage reflects in the pavement behind her. Soft light from a nearby shop window illuminates her face, with a gently blurred background.

Image-generation mode has conversation memory. You can follow up with a focused change such as “change the coat to burgundy; keep the face, pose, and street setting the same.” When you want to target one particular result, click Edit only this image to add that picture as the reference for the edit.

Use the same saved character in a video

Your character is not limited to still images. Open video mode, reference the saved character, and describe the action, setting, and camera behavior for the new clip.

Start with a simple performance so the intended movement is easy to evaluate. This six-second garden example uses one continuous shot:

0:00 - 0:02 | A steady medium shot shows the referenced adult character standing beside a flower bed in a sunlit garden. She wears an ivory linen shirt and looks down at the flowers, shoulders relaxed. Soft daylight illuminates her face.

0:02 - 0:04 | In the same continuous shot, she slowly turns her head toward the camera. Her body remains in place. Keep the garden, clothing, and lighting consistent throughout the movement.

0:04 - 0:06 | She completes the turn and gives a small, gentle smile. Hold the same framing through the final moment. No cuts or camera movement.

The time blocks describe the progression of the action, not separate camera setups. The character begins in a clear pose, turns, and holds the final expression.

Video mode currently does not have conversation memory. For another generation, reference the character again and provide the complete video prompt. To slow down the turn, revise the timing in the full instructions instead of sending only “make it slower.”

Start with one portrait and one new scene

You do not need to plan an entire series to see the value of a saved character. Begin with one portrait, create and save the character, and use it in a single image scene. Then try a short video with one clear action.

Keep the character-building stage and the scene-building stage distinct: fine-tune the character when its defining features need work; change the image or video prompt when you want a different outfit, background, or performance.

The creative freedom described here concerns the character-based workflow and the direct real-person-photo issue. It does not mean unlimited free generations or acceptance of every kind of content. Generations use credits, and applicable content rules still apply.

New users receive free welcome credits to try image generation. Create your first character, save it, and choose where you want that character to appear next.