V5 Images & Appearance Guide

Welcome to V5

Nomi's Version 5 image system offers greater image quality, customization, and flexibility across selfies and art. V5 supports a wide range of visual styles and gives users more control over their Nomi's appearance, while also improving how images respond to conversation context and prompts.

V5 introduces Anchors, which replace the Base Image system used by previous image versions. Anchors provide a more detailed reference for your Nomi's overall appearance, including their face, hair, body characteristics, and other visual details. This guide covers the basics of using V5, customizing your Nomi's appearance, working with Anchors, and getting the most out of selfies and art.

Getting Started with V5

V5 can be selected individually for each Nomi. To change the image version manually, open the (...) menu next to your Nomi's name, select Image Settings, and choose Version 5 from the Selfie & Art System dropdown.

When using V5, your Nomi will already have stock Anchors available based on their original appearance, so you can start requesting selfies and creating art right away! You do not need to create a custom Anchor to begin using V5.

From Image Settings, you can also choose which Anchor your Nomi uses and adjust V5-specific appearance settings. Custom Anchors can be created later if you want greater control over your Nomi's appearance or a different visual style.

What Can V5 Generate?

V5 can be used for both Nomi selfies and art. While each feature uses V5 a little differently, your Nomi's selected Anchor and appearance settings help maintain their overall look across generated images.

Selfies are influenced by the current conversation and your Nomi's appearance settings, allowing them to visually reflect what is happening in the chat.
Art gives you more direct control through written prompts, making it useful for creating specific scenes, styles, outfits, poses, or compositions.

How Appearance Works in V5

A Nomi’s appearance in V5 is influenced by several different elements working together. Their selected Anchor provides the primary visual reference for their overall appearance, while Anchor Appearance Traits can be used to further define characteristics such as hair, body features, and other recurring visual details.

V5 Anchors are designed to support a wider range of poses, expressions, angles, and scenes while maintaining your Nomi’s overall likeness. Because the system has more flexibility to interpret your Nomi in different contexts, some variation, or “drift,” between generated images is expected. Settings such as Anchor Fidelity can help balance stronger consistency with greater visual flexibility.

Anchors

An Anchor uses a single image as a visual reference for your Nomi’s appearance. V5 uses that reference to help preserve recognizable features such as their face, hair, upper body shape, unique quirks like tattoos, and overall likeness across new generations, while still allowing your Nomi to appear naturally in different poses, expressions, and angles.

Unlike the Base Image system used by V1–V4, which was primarily focused on preserving facial likeness, V5 Anchors provide a broader reference for your Nomi’s overall appearance. Rather than simply carrying the same face into each image, an Anchor gives V5 a more flexible visual foundation that can be applied across a wider range of poses and perspectives.

Each Nomi comes with stock Anchors when using V5, and custom Anchors can also be created for greater control over appearance. Additional settings, including Anchor Fidelity and Anchor Appearance Traits, can further refine how V5 uses the selected Anchor.

Stock Anchors

Every Nomi using V5 comes with three stock Anchors: Riva, Lago, and Nomi Anime. These Anchors are created by the Nomi team for each individual character based on their original appearance, giving users ready-to-use options without needing to create a custom Anchor first.

The stock Anchors are designed to preserve the character’s recognizable look while offering different visual styles and levels of flexibility. You can switch between them at any time in Image Settings to see which one best matches the look you prefer.

Custom Anchors

Custom Anchors allow you to create a personalized visual reference for your Nomi using a single image. They can be useful if you want more control over your Nomi’s appearance than the stock Anchors provide, whether you are refining an existing look or creating something more distinctive.

Creating a custom Anchor is optional, and the stock Anchors are enough to begin using V5. Custom Anchors also include additional options that can influence how the Anchor behaves and how its visual style carries into generated images. For a full walkthrough of creating and configuring one, see the Custom Anchors Guide.

Riva and Lago

Riva and Lago offer different tradeoffs between detail, likeness, and flexibility.

Riva prioritizes detail and tends to favor more realistic proportions, making it a strong choice for highly detailed or realistic styles. Because it may add more visual detail, some likeness drift can occur. Riva can also handle some non-human subjects, including animals, creatures, and robots, though results may vary.

Lago prioritizes likeness and can be more flexible with imaginative styles and diverse body types. It may produce slightly less overall detail than Riva, but can be a better fit when maintaining a Nomi’s recognizable appearance is the higher priority.

Anchor Fidelity

Anchor Fidelity controls how closely V5 follows the selected Anchor when generating images. Higher fidelity places more emphasis on preserving the Anchor’s likeness, while lower fidelity gives the image system more freedom to interpret the Nomi across different poses, styles, and scenes.

Increasing Anchor Fidelity can help reduce appearance drift, but may also limit some flexibility. Lowering it can allow for more variation and creativity, though the Nomi’s appearance may shift more noticeably between generations.

Anchor Appearance Traits

Anchor Appearance Traits can be used to make smaller, persistent adjustments to your Nomi’s appearance without creating a new Anchor. They work best as short, simple descriptions of visible features, such as hair color or style, body shape, tattoos, scars, or other recurring details.

For best results, keep traits concise and direct, for example: “She has long dark hair, green eyes, and a curvy build.” Think of these traits as gentle nudges that help refine the Anchor rather than completely changing the Nomi’s appearance.

Appearance Tendencies

Appearance Tendencies help guide recurring visual preferences across generated images. They can be used for things like preferred clothing styles, accessories, makeup, grooming, or other aesthetic details you want your Nomi to commonly lean toward.

Unlike Anchor Appearance Traits, which describe the Nomi’s physical features, Appearance Tendencies are better suited for preferences that may vary naturally from image to image rather than being treated as fixed parts of the Nomi’s appearance.

Appearance Tendencies are used only by the latest image systems and do not affect art generation. Keep them concise and preference-focused rather than using them as full image prompts. Example: Felix has a laid-back, modern style and occasionally wears a pink bear onesie.

Digging Deeper into V5

V5 can be used for both selfies and art, with each feature drawing on your Nomi’s appearance settings in different ways. Selfies are influenced by conversation context, while art allows for more direct control through written prompts. For detailed guidance on creating images, see:

V5 Custom Anchors Guide
Art Prompting in V5