This catalogue discusses fictional adult AI companions and is intended for visitors aged 18 or older. The workshop notes describe what to check on partner platforms, not a live service hosted here. By continuing, you confirm you are an adult and understand that any companion interaction happens on the partner's terms.
This site is an independent editorial catalogue. No chat runs on this site; partner links lead to third-party platforms with their own age, privacy and billing rules.
Dondi AI is best understood as a bench, not a shelf. You start with a base skin, then hang styling, tone, voice and backstory on top like separate coats. Each layer can be pulled back, swapped or left alone. The workshop metaphor matters because companion pages usually show only the finished figure. Here we look at the seams: what is fixed at creation, what can be reworked later, and which layers tend to become expensive to change. The goal is not a perfect build on day one, but a persona you can keep adjusting without starting over.
Treat identity as composable layers: a skin is not just looks, it is appearance plus tone plus voice plus history, and each layer has its own cost and reversibility.
Worth your time if you enjoy slowly refining one character and can tolerate a slower, more deliberate build. Each added layer gives the persona more grip, and the workshop format rewards curiosity about how appearance, tone and history interact. A poor fit if you collect companions like postcards and want a different face every week; depth here is a long game, and layer edits can cost more than a fresh start. Treat Dondi as a workbench, not a vending machine, and verify what each change costs on the partner page before you commit.
Finished companion galleries hide the joins. A workshop exposes them. In this catalogue we separate a Dondi persona into six coats: base skin, wardrobe, tone, voice, backstory and portability. Each coat has its own reversibility, its own likely cost and its own effect on how the character feels after the first week. Some layers are structural, like the base; others are trim, like an outfit set. Knowing which is which keeps you from rebuilding the whole bench when you only wanted a different jacket. The six cards below are the checks we would run before adding anything permanent.
01
Base skin and visual identity
The base is the silhouette the rest of the persona hangs on: face shape, build, palette and the general visual register. In a workshop build, the useful question is not whether the base looks good in the preview, but how much of it stays adjustable after registration. Some platforms lock the base entirely; others allow partial swaps at a cost. A base that cannot be nudged becomes a permanent compromise in every future image, which is why we treat it as the least flexible layer. Decide what you could live with for months, not what looks best in a single still.
Every later layer sits on top of this one, so an immovable base quietly sets the ceiling for the whole persona. If the silhouette is wrong, wardrobe and lighting become patches rather than improvements.
What to check first: Ask the partner whether the base can be edited later, how that edit is billed, and whether swapping the base resets wardrobe, tone or history. Confirm this before you invest in detail.
02
Wardrobe and styling layers
Clothing, hair and accessories are the most visibly composable part of a persona. In a good workshop they behave like saved sets rather than one-off descriptions: you define a look once and reuse it across future images. That matters because consistency is expensive to recreate from scratch every session. The tradeoff is that saved styling narrows variety, and a wardrobe that works in one context can feel repetitive in another. We look for whether styles are stored, whether they survive edits to other layers, and whether the platform treats a new outfit as a fresh creation or a simple toggle.
Saved styling is the difference between a character who looks continuous and one who needs re-explaining every session. It also cuts the number of separate builds you pay for when you only wanted a different jacket.
What to check first: Check whether saved outfits apply to future images automatically, whether they survive edits to tone or voice, and whether changing an outfit counts as a new build on your tier.
03
Tone presets and custom voice
Tone is how the persona speaks in text: warm, dry, teasing, formal, or something you describe yourself. Presets are convenient but tend to flatten over long conversations, drifting toward a generic helpfulness. A free-text tone field lets you write the register in your own words and correct it later, which usually holds longer. This layer is invisible in a gallery thumbnail but it is the one people return for. We treat tone as structural rather than cosmetic, because it shapes how every other layer is experienced. Changing tone after a long build can feel like meeting a different person.
Conversation quality lives here. Appearance gets attention, but tone decides whether you come back tomorrow, and it is harder to retrofit than any outfit. A persona with a weak tone feels like a mannequin wearing good clothes.
What to check first: Confirm that tone can be rewritten in your own words after creation, not just chosen from fixed presets, and test whether edits apply to an existing conversation or reset the persona.
04
Audible voice casting
If the companion speaks aloud, voice becomes a layer you will likely want to revisit. A workshop should offer previews before commitment, because a written description of a voice rarely matches the felt impression. We look at how many voices are available, whether they are sampled in context, and whether recasting later is treated as a small edit or a full rebuild. Voice is often the layer people change after a week, once they know how the persona actually behaves. A good casting decision in week one can save a second creation fee in week three.
Voice is the layer most likely to be reconsidered after you have lived with the persona for a while. Hearing a sample in isolation is not the same as hearing it in dialogue.
What to check first: Preview voices in actual conversation if possible, ask whether recasting later costs the same as the initial casting, and confirm whether a voice change affects the rest of the persona.
05
Backstory as the persona spine
Backstory is the spine that keeps the other layers from drifting into generic territory. It can be a few lines or a long history, but what matters is whether the platform actually reads it during conversation. A workshop approach means testing the spine with a question only that history could answer, rather than trusting that a long field is being used. Backstory also influences tone and voice indirectly, because a persona with a past has reasons to speak the way it does. We treat it as the load-bearing layer: thin here and everything above feels decorative.
Without a spine, wardrobe and tone become a costume rather than a character with reasons. A backstory that is never referenced is just a longer description field, and it is the layer most likely to be silently ignored.
What to check first: Test with a backstory-specific question before assuming the history is active, check whether the field has a practical length limit, and see whether edits to history ripple into tone.
06
Persona portability and reuse
Portability is whether a finished workshop can become a template: duplicating a persona, saving it as a starting point, or moving its layers onto another base. This is the layer that turns a slow first build into a fast second one. Platforms vary widely here; some charge full creation cost for a duplicate, others allow cheap variation. We look for whether duplication preserves tone and history or resets them, and whether a template flattens the differences you worked to create. A reusable persona is a tool; a locked one is a one-off sculpture. Know which you are buying.
Reuse changes the economics of a second build, but templates can sand away the details that made the first one yours. The question is whether speed is worth the loss of specificity.
What to check first: Verify duplication behaviour before you rely on it: ask whether a duplicate is billed as a fresh creation, and whether tone and history carry over or reset.
Six partner paths
Compare benches
The comparison page is not a scoreboard. It is a set of benches: each partner is checked against the same six persona layers so you can see where a given platform might reward depth and where it might charge for changes. We do not rank brands or claim fixed capabilities. Use the layer questions to compare how easily appearance, tone, voice and history can be adjusted, then verify current behaviour with each partner before you register.
The catalogue lists companion platforms by name only, without endorsing specific features. Each entry points to the partner's own description; the workshop layers remain our editorial lens. If a platform is listed, it does not mean every layer is editable or affordable. Treat the catalogue as a starting shelf, then apply the six checks to whichever bench you want to work at. Confirm current availability, editing rules and prices on the partner.
Start with the silhouette you could tolerate for months, not the one that looks best in a single frame. The base is the least reversible layer in most workshop builds, and a poor choice here forces every later layer to compensate. If the platform allows a base swap later, find out what it costs and whether it resets wardrobe or tone. If it does not, choose as if you were signing a lease. This is the slowest decision and the one most people rush.
Apply tone before wardrobe
Set the conversational register early, because tone shapes how the persona develops and is harder to retrofit than clothing. Write the register in your own words if the platform allows it, and test it across several exchanges before adding visual detail. A persona with the right tone can survive a mediocre outfit; the reverse is rarely true. This order also keeps you from paying for visual generations that do not match the character you are actually building toward.
Cast the voice last
Choose an audible voice only after the visual and tonal identity is settled. Voice casting is the layer most likely to be regretted, and previews in isolation can mislead. If you can hear samples in dialogue, do that. Otherwise, start with the closest match and plan for a possible recast. Confirm what a voice change costs before you rely on it later. Casting last means the voice is chosen for a finished persona rather than a rough sketch that may still shift.
Register the persona on the partner page
Take the finished stack to the partner page and read what is actually offered today. Confirm what editing costs, whether personas can be duplicated, how voice changes are billed, and whether tone fields accept free text. Availability and terms belong to the partner and can change, so treat any workshop note as a question rather than a promise. This is also where you decide whether the first-year figure is worth it for the depth you intend to use. Register only after the checks match your expectations.
What Each Layer Costs You
Need
Look for
Trade-off
I want one companion refined over time
Editable layers and saved styling sets
Layer depth rewards patience and rarely pays off for casual, short-term use
I want visual consistency across sessions
Saved wardrobe or style sets that reuse cleanly
Consistency reduces variety and can make repeated images feel similar
I want a specific conversational register
Free-text tone fields rather than fixed presets
Custom tone needs occasional correction as the conversation develops
I want a voice I can change my mind about
Previewable voices and low-cost or free recasting
Voice changes are often billed as a fresh creation
I want genuine history behind the persona
Long backstory fields that visibly shape replies
A long description does not guarantee long context or active use
I want to build a second persona quickly
Duplication or templates from an existing build
Templates flatten the differences you worked to create
Workshop Notes
This is an independent editorial blueprint for this Persona Workshop catalogue. Creation, editing, limits and costs belong to the partner and can change without notice. Treat every layer note as a question to verify on the partner page, not as a fixed capability.
Persona depth is a framework, not a performance claim. Separating appearance, tone, voice and history does not guarantee a better conversation or a more satisfying companion. It simply gives you language to ask better questions before spending.
Every character referenced here is a fictional adult. Nothing on this page describes real people, and no on-site chat exists on this site; any conversation happens on the partner platform under its own terms.
Reversibility is not guaranteed on any layer. Some edits may be treated as a new creation and billed accordingly. Confirm the current rules for base skins, styling, tone, voice and duplication before investing in a detailed build.
We may earn a commission when visitors sign up through partner links. That does not change the workshop framework, but it does mean you should verify current availability, prices and privacy practices with the partner before registering.
Frequently asked questions
What is a persona skin in this workshop?
A skin here is the full stack that defines a companion: the base appearance, styling, conversational tone, audible voice and backstory. A workshop treats those as separate coats rather than one indivisible character. That matters because each coat has its own reversibility and cost. You might be able to change an outfit cheaply while a base swap resets everything. Thinking in skins helps you ask what is actually editable before you build something elaborate, instead of discovering the joins after you have already invested time and balance.
Can I undo a layer I regret?
Sometimes, but not uniformly. Reversibility varies by platform and by layer, and some changes may be treated as a new creation. A wardrobe toggle is often cheap; a voice recast may not be. The workshop view is to ask before building: which layers can be edited, what each edit costs, and whether an edit resets adjacent layers. If the partner does not make this clear, assume the change is expensive and plan accordingly. Do not rely on undo as a design strategy.
Is tone or appearance more important?
Appearance gets you to the first conversation; tone is what keeps you coming back. A persona with the right conversational register can survive a mediocre outfit, while the reverse is rarely true. In a workshop build we would spend early effort on tone, then wardrobe, then voice. That order also prevents expensive visual generations from being wasted on a character whose personality is still undefined. If you can only invest in one layer, put your attention into tone and test it across several exchanges.
Should I duplicate a persona I like?
Only if duplication is cheap on your tier. Otherwise a small variation on the original is usually faster than rebuilding the whole stack. Ask whether a duplicate carries over tone, voice and history, or resets them; some platforms bill a duplicate as a fresh creation with full layer costs. If duplication is expensive, consider whether you actually want a second persona or just a different outfit for the first. The workshop metaphor is useful here: a template is only worth it if it saves more time than it costs.
Does a long backstory improve conversation?
It can, if the platform actually uses it. A long backstory field is not proof of long context or active recall. Test it with a question only your history could answer, and see whether the reply reflects what you wrote. If the answer is generic, the spine is decorative. Some platforms may have practical limits on how much history is read. Write the backstory as if it matters, but verify that it lands before you build more layers on top of it.
Why does this page avoid direct partner comparisons?
Because layer behaviour changes faster than editorial pages do. A comparison that is accurate today may be misleading after a partner updates its editing rules, voice casting or pricing. The workshop framework stays useful because it is about what to check, not which brand wins. You can apply the six layers to any companion platform and ask better questions. We would rather give you a durable checklist than a ranking that ages badly. Verify current specifics with each partner before you commit.
