---
title: "Consistent characters, mascots and products"
description: "The consistency problem Generating one great image is easy. Generating twenty images where the same mascot, model or product looks identical — same face…"
url: https://optimizeall.com/learn/ai-image-generation-and-design/consistent-characters-and-products
updated: 2026-10-05
---

AI Image Generation and Design · Consistency across a campaign · lesson 10 of 18 · 8 min

# Consistent characters, mascots and products

## The consistency problem

Generating one great image is easy. Generating twenty images where the same mascot, model or product looks identical — same face, outfit, proportions, colors — is much harder, because each generation starts from new random noise. Campaigns, carousels, storyboards and brand mascots all depend on solving this.

## Techniques from simplest to most advanced

| Technique | How it works | Strength | Limits |
|---|---|---|---|
| Detailed descriptive anchor | Same precise description in every prompt | Easy | Drift in details |
| Fixed seed + small prompt changes | Same starting noise | Helpful for close variations | Breaks with larger changes/model updates |
| Character/subject reference | Upload approved image(s) of the character | Strong for faces/outfits | Tool-dependent quality |
| Character sheet | Generate or draw a reference sheet (front, side, expressions) | Gives the model and team a canonical look | Needs curation |
| Image-to-image from approved frames | Use an approved image as a base | Keeps composition/look | Can reduce variety |
| Custom fine-tuning / adapters | Train a small model add-on on approved images (e.g. LoRA-style techniques) | Very consistent | Technical skill, rights to training images, platform support |
| Compositing real product photos | Use real photos for the product | Perfect accuracy | Requires matching light |

## The character bible

For mascots and recurring characters, document everything:

```
Character bible: "Zara the Zebra" (brand mascot)
Canonical description: a friendly cartoon zebra with rounded features,
  large expressive eyes, teal scarf, black-and-white stripes (5 across the body)
Proportions:  head ~1/3 of body height
Style:        flat vector, soft shading, brand palette
Expressions:  happy, curious, surprised, thinking (reference images)
Poses:        standing, waving, pointing, sitting
Do:           teal scarf always; stripes consistent
Don't:        realistic fur; extra accessories; other colors on scarf
Reference files: zara_front.png, zara_side.png, zara_expressions.png
```

Consider having an illustrator create the canonical mascot design by hand. A human-designed mascot is easier to protect as intellectual property and gives you a clean master to reference.

## Real people and consistency

Creating consistent AI versions of **real people** — yourself, a spokesperson, a model — requires explicit consent from that person, clear agreement on how their likeness will be used, and care about platform policies on synthetic media and impersonation. Many tools restrict generating likenesses of real people, especially public figures. For consistent synthetic "people" who do not exist, avoid presenting them as real customers, employees or reviewers.

## Products: accuracy first

For products, accuracy usually matters more than creativity:

- Use real product photography as the source of truth.
- Use product-reference features only if outputs preserve logo, shape, color and label exactly — check each image.
- When in doubt, composite the real product into AI scenes.
- Never alter a product's appearance in ways customers would find misleading.

## Keeping settings stable

Model updates change outputs. For a campaign:

- Note the tool and model version in your prompt log.
- Produce campaign assets within a defined period if possible.
- Keep approved images as the master references — not just prompts.

## Worked example: a children's education creator

A creator uses a mascot owl in weekly carousels.

1. An illustrator designs the owl by hand; copyright is assigned to the creator.
2. A character sheet (front, side, six expressions) is created.
3. The creator uses the sheet images as character references and a style reference at medium strength.
4. Each weekly prompt uses the canonical description plus the scene.
5. Outputs are checked against the bible (colors, proportions, accessories); small deviations fixed by inpainting.
6. After six months, the approved outputs are curated as additional references.

## Common mistakes

- Expecting a text description alone to keep a character identical.
- Using a real person's likeness without consent.
- Letting product details drift (wrong logo, color or label).
- Relying on prompts only, without saving approved images as references.

## Checking consistency efficiently

Consistency checks are faster with side-by-side comparison. Place each new output next to the canonical reference images and compare a short list of features: face shape and proportions, colors of key clothing or product parts, distinctive marks (such as the mascot's scarf or the product's label), and style (line weight, shading, texture). A simple grid of all approved images for a campaign also reveals drift that is invisible when images are reviewed one at a time.

## What's new: multi-reference models

Consistency got much easier in 2025–2026 because several models now accept multiple reference images in one request and are designed to hold identity across scenes. Google, for example, says Gemini 3 Pro Image can use up to 14 reference images and keep up to five people consistent; Ideogram offers a character reference feature; Midjourney and Adobe Firefly offer their own reference tools. The workflow is the same everywhere: **approved references in, one controlled change out, human check at 100%.**

## Hands-on: a reference pack and a consistency prompt

```text
REFERENCE PACK — "Zara the Zebra"
zara_front.png      zara_side.png       zara_34.png
zara_expressions.png (happy, curious, surprised, thinking)
zara_scarf_detail.png   palette_card.png

CONSISTENCY PROMPT (assistant-built model with multiple references)
"Use the attached reference images of Zara as the only source for her
design: same stripe count (5 across the body), same teal scarf, same
proportions (head about one third of body height), same flat vector style.
New scene: Zara reading a picture book under a tree in a park, curious
expression, soft morning light, 4:5, space at the top for a title.
Do not add accessories, text or logos."
```

## Consistency scorecard

Check every new image against the bible before it enters the approved set:

| Attribute | Canonical | Image 7 | Pass? |
|---|---|---|---|
| Stripe count on body | 5 | 6 | Fix (inpaint) |
| Scarf color | Brand teal | Teal | Pass |
| Head-to-body ratio | ~1/3 | ~1/3 | Pass |
| Eye shape | Large, rounded | Large, rounded | Pass |
| Style | Flat vector, soft shading | Flat vector | Pass |

Only images that pass go into the reference folder; otherwise, errors compound — a slightly wrong image used as a reference makes the next one more wrong.

## Real people: the consent line

Consistent "characters" based on real people — founders, employees, influencers, customers — need **written, specific consent** covering AI generation, uses, duration and withdrawal. Several platforms and jurisdictions also treat realistic synthetic depictions of real people as content that must be labeled. When in doubt, shoot the real person or design a fictional character.

## Summary

Combine descriptive anchors, references, character sheets and — where appropriate — fine-tuning or compositing to keep subjects consistent; document a character bible; get consent for real likenesses; and treat real product photos as the source of truth.

## Video lecture: Consistent characters, mascots and products

Lecture coming soon · 10 chapters · about 8 minutes. Read the full transcript below.

1. Consistent characters and products
2. Why consistency is hard
3. Techniques, simple → advanced
4. The character bible
5. Worked example 1: Zara, weekly posts
6. Worked example 2: six juice flavors
7. Watch me do it: a reference pack
8. Real people: the consent line
9. Common mistakes
10. Recap and try this now

## Lecture transcript

### Consistent characters and products

Making one great image of a mascot is easy. Making twenty images where she has the same face, the same scarf, the same five stripes and the same proportions? That's where most AI campaigns fall apart. By image seven, your friendly zebra has six stripes, a slightly different nose, and suddenly a brooch nobody asked for. Followers notice. Kids especially notice. In this lecture, you'll learn the consistency techniques from simplest to most advanced, how to build a character bible and a reference pack, how to keep real products accurate, and where the consent line sits for real people. By the end, you'll be able to run a recurring character or product across a whole campaign without drift.

### Why consistency is hard

Why is this hard? Remember that each generation starts from new random noise. Nothing in the model remembers what your character looked like last time unless you tell it, every time, with words and pictures. And small errors compound. If a slightly wrong image becomes a reference for the next one, the next one gets more wrong. It's like a game of telephone. Each whisper changes the message a little, and by the end of the line, it's a different sentence. Consistency matters commercially too. Mascots build recognition only when they're stable, and products must match what customers actually receive. An inconsistent product image isn't a style issue. It's an accuracy problem.

### Techniques, simple → advanced

Here are the techniques, from simplest to most advanced. A detailed descriptive anchor: the same precise description in every prompt. A fixed seed with small changes, useful for close variations. Character or subject references: uploading approved images. A character sheet with front, side, three-quarter and expressions, which gives both the model and your team a canonical look. Image to image from approved frames. Custom fine-tuning, like small adapter models trained on approved images, for very high consistency. And for products, compositing real product photos, which gives perfect accuracy. The newest shift is multi-reference models. Several assistant-built models now accept many reference images at once and are designed to hold identity across scenes.

### The character bible

The heart of the process is the character bible. It's a one-page document with a canonical description, proportions, palette, style, expressions, wardrobe rules and do's and don'ts. For Zara the zebra: a friendly cartoon zebra, rounded features, large expressive eyes, a teal scarf, exactly five stripes across the body, head about a third of body height, flat vector style. Alongside it sits the reference pack: front, side and three-quarter views, an expressions sheet, a scarf detail, and a palette card. Here's the key idea. The bible and the reference pack are the source of truth, not the last image you generated. Every new image is checked against them before it's approved.

### Worked example 1: Zara, weekly posts

Worked example one, simple. A children's education creator uses Zara in weekly storybook posts. She builds the bible and reference pack. For each new scene, she attaches the references and writes a consistency prompt: use the attached references as the only source for her design, same stripe count, same scarf, same proportions, same style. Then the new scene, Zara reading under a tree. Image seven comes back with six stripes. Instead of regenerating, she inpaints one stripe away, then checks the consistency scorecard: stripes, scarf color, head ratio, eye shape, style. All pass. Only then does image seven join the approved folder. That one rule stops the telephone game.

### Worked example 2: six juice flavors

Worked example two, products, with illustrative details. An agency in Lahore produces a monthly campaign for a juice brand with six bottle flavors. Early tests show the model changes the label text, the cap color and the bottle's curve. So the agency makes a rule. The product is never generated. They photograph each bottle from three angles on a neutral background: front, three-quarter and side, with consistent lighting. The AI only generates scenes, leaving room for the bottle in the right position and angle. Then they composite the real bottle, add contact shadows, and match the grade. Result: every scene is new, and every bottle is exactly right. The client's regulatory team approves labels once, not every month.

### Watch me do it: a reference pack

Watch me do it. Let me build a reference pack for a mascot from scratch. First, I create the canonical front view with the illustrator, or pick the best approved generation and clean it up by hand. Next, I generate side and three-quarter views using that front view as the reference, and I fix details manually until all three agree. Then an expressions sheet, happy, curious, surprised, thinking, all from the same references. I add a palette card with the exact brand colors. I save everything with clear names in one folder, and write the bible. Finally, I run a test: three new scenes, checked with the scorecard. If anything fails twice, I improve the references, not the prompt.

### Real people: the consent line

Now real people, because consistency techniques work on faces too. A consistent AI version of a founder, an employee, an influencer or a customer needs written, specific consent. That means covering AI generation, where the images will be used, for how long, and how they can withdraw. Several platforms require labels on realistic synthetic depictions of people, and laws on likeness and personality rights vary by country. The EU AI Act's transparency rules, for example, require deployers to disclose deepfakes, including realistic images of real people, from August twenty twenty-six. When in doubt, either photograph the real person or design a clearly fictional character. Fictional characters are also easier to own and protect.

### Common mistakes

Common mistakes. Starting production before the bible exists. Using drifted images as new references, which compounds errors. Describing the character differently in each prompt. Letting the model redesign the product label or the logo. Trusting a model update to keep your character the same, when new versions can interpret references differently. And skipping the scorecard when you're in a hurry. One more. Don't forget settings. If your mascot lives in a recurring place, like a classroom or a shop, make a setting reference too. Consistent places make consistent characters feel real, and they save you prompt space.

### Recap and try this now

Let's recap. Consistency doesn't happen by itself, because each generation starts fresh. Use a character bible and a reference pack as the source of truth, write a consistency prompt that names the fixed attributes, check every image with a scorecard, and only approved images become references. For products, composite real photos so labels and shapes stay exact. For real people, get written consent and label realistic depictions. Try this now. Create a bible for one recurring subject, a mascot, a character or a product, with at least three reference images and a do and don't list. Then generate three new scenes and score each one before you approve it.

## Key takeaways

- Consistency must be engineered: every generation starts fresh, and errors compound when drifted images become references.
- A character bible and curated reference pack are the source of truth; check every new image with a scorecard.
- Multi-reference models help, but approved references and human checks remain essential.
- Composite real product photos so labels, caps and shapes stay exactly right.
- Consistent depictions of real people require written, specific consent and appropriate labeling.

## Try it

Create a character or product bible for a recurring subject, including a canonical description, do/don't rules and at least three reference images, then generate three scenes and check them against it.

- [Previous: Compositing AI output with photography and typography](https://optimizeall.com/learn/ai-image-generation-and-design/compositing-ai-with-design)
- [Next: Style systems and campaign prompt libraries](https://optimizeall.com/learn/ai-image-generation-and-design/style-systems-and-prompt-libraries)
- [All lessons of AI Image Generation and Design](https://optimizeall.com/learn/ai-image-generation-and-design)