Private models and dedicated capacity

Custom Furry AI Models and a Dedicated API

Our plans already go to unlimited images, so a bigger quota is not what you are missing. This is a model trained on your characters, drawing them the same way in every render, in stills and in video. Private LoRA, monthly retuning, your own endpoint. Adult furry work allowed.

Request a custom plan See standard plans

Made on the same stack 🐾

Public generations, unretouched. A private model starts here, then learns your characters.

What a custom plan buys

Four things no quota can buy.

A model that is yours

Trained on your references. Scoped to your account, never merged into the public models.

Characters that stay themselves

Same fursona across scenes, poses and outfits. A prompt describes; a trained model reproduces.

Retuning and isolation

New references folded in monthly. Your data stays out of the public pool and the dataset.

Capacity you do not share

Your own endpoint on provisioned GPU. Concurrency agreed up front, not a shared queue.

How private model loading works

1. You send references. Art, sheets, past generations, or an existing LoRA. We say what is usable before you are billed.
2. We prepare the dataset. Cleaned, captioned in e621 tags, near duplicates cut.
3. We train, you review. You get a test sheet. Character off means we retrain, inside the setup fee.
4. It loads privately. Bound to your account, selectable in the generator or callable from your API key. Never in the public checkpoint list.
5. It gets versioned. Each retune is a new version, the old one stays. A bad retune is a rollback, not a loss.

Your references and outputs are never exported to the training dataset. Ask us to delete a model and the training set goes with it.

Three ways to buy

Setup and monthly are quoted separately. Entry prices below; the quote follows the number of characters, the retuning cadence and the capacity you need.

Creator Custom

For a creator with characters that have to stay themselves.

From $900 setup

then from $500 a month

One or more private LoRAs trained on your characters
Character consistency across scenes, poses and outfits
Presets we validate with you, so a good render is repeatable
Generous image and video allowance, priority in the queue
Retuning on an agreed cadence as your world grows
Commercial rights written down

Studio Video

For a studio or AI-influencer account shipping clips on a schedule.

From $1,500 setup

then from $1,200 a month

Your character model, animated: stills stay on-model in motion
300 video credits a month included, 720p, 5s or 10s, with audio
Batch runs from a still sheet, not one clip at a time
Overage billed per clip, agreed up front
Retention set by you, 180 days by default

Dedicated API

For a product or studio generating on its own schedule.

From $2,500 setup

then from $2,000 a month

Your own endpoint and key, isolated from the public site
Included volume with agreed overage, no surprise throttling
Guaranteed concurrency on provisioned GPU capacity
Private models callable straight from the API
Retention and deletion rules set by you
Named contact and a response-time commitment

Against a standard plan

Only the things a quota cannot fix.

Feature	Custom	Standard plans
Private model trained on your characters Standard plans share the public checkpoints.
Same character across every render A prompt describes. A trained model reproduces.	Trained in	Prompt luck
Monthly retuning as your world grows
Your references and outputs kept out of the public pool
API access Standard plans are web only.	Dedicated endpoint
Throughput Own GPU, never queued behind the public site.	Provisioned per contract	Shared queue
Your character in video Standard plans animate any still. A private model keeps the character on-model.	Animated from your model	Generic i2v
Adult and furry content allowed
Billing Neither auto-charges. We keep no card.	Invoiced per period	Self-serve, one payment

Request a custom plan

What needs to stay consistent, and roughly what you expect to generate. The volume figure sets the capacity we quote. We answer within one business day.

Frequently asked questions

Who owns the model you train for me?: You do. It stays scoped to your account and is never merged into the public checkpoints. Ask us to delete it and the training set goes with it.
Do my references or generations train anything else?: No. Nothing you send and nothing your private model produces is exported to the dataset.
Is adult furry content actually allowed?: Yes. Same limits as the public site: everyone depicted is adult and fictional, and the prompt filter still applies.
How long before it is running?: About a week for a first Creator Custom model. Two to three for a dedicated API.
What exactly is a clip?: 720p, 5 or 10 seconds, with audio, animated from a still. It is image-to-video, so the still comes from your model first, which is what keeps the character on-model.
How does billing work?: Setup once, then monthly, invoiced per period. We hold no card on file, so nothing is ever charged automatically.
What if I only need a bigger quota?: Then skip this. Standard plans already go to unlimited images.

Bring us a character that has to stay itself

A model that knows your characters, and capacity that is yours.

Request a custom plan See standard plans

Custom Furry AI Models and a Dedicated API

What a custom plan buys

Four things no quota can buy.

A model that is yours

Trained on your references. Scoped to your account, never merged into the public models.

Characters that stay themselves

Same fursona across scenes, poses and outfits. A prompt describes; a trained model reproduces.

Retuning and isolation

New references folded in monthly. Your data stays out of the public pool and the dataset.

Capacity you do not share

Your own endpoint on provisioned GPU. Concurrency agreed up front, not a shared queue.

How private model loading works

1. You send references. Art, sheets, past generations, or an existing LoRA. We say what is usable before you are billed.
2. We prepare the dataset. Cleaned, captioned in e621 tags, near duplicates cut.
3. We train, you review. You get a test sheet. Character off means we retrain, inside the setup fee.
4. It loads privately. Bound to your account, selectable in the generator or callable from your API key. Never in the public checkpoint list.
5. It gets versioned. Each retune is a new version, the old one stays. A bad retune is a rollback, not a loss.

Your references and outputs are never exported to the training dataset. Ask us to delete a model and the training set goes with it.

Three ways to buy

Setup and monthly are quoted separately. Entry prices below; the quote follows the number of characters, the retuning cadence and the capacity you need.

Creator Custom

For a creator with characters that have to stay themselves.

From $900 setup

then from $500 a month

One or more private LoRAs trained on your characters
Character consistency across scenes, poses and outfits
Presets we validate with you, so a good render is repeatable
Generous image and video allowance, priority in the queue
Retuning on an agreed cadence as your world grows
Commercial rights written down

Studio Video

For a studio or AI-influencer account shipping clips on a schedule.

From $1,500 setup

then from $1,200 a month

Your character model, animated: stills stay on-model in motion
300 video credits a month included, 720p, 5s or 10s, with audio
Batch runs from a still sheet, not one clip at a time
Overage billed per clip, agreed up front
Retention set by you, 180 days by default

Dedicated API

For a product or studio generating on its own schedule.

From $2,500 setup

then from $2,000 a month

Your own endpoint and key, isolated from the public site
Included volume with agreed overage, no surprise throttling
Guaranteed concurrency on provisioned GPU capacity
Private models callable straight from the API
Retention and deletion rules set by you
Named contact and a response-time commitment

Feature

Custom

Standard plans

Private model trained on your characters

Standard plans share the public checkpoints.

Same character across every render

A prompt describes. A trained model reproduces.

Trained in

Prompt luck

Monthly retuning as your world grows

Your references and outputs kept out of the public pool

API access

Standard plans are web only.

Dedicated endpoint

Throughput

Own GPU, never queued behind the public site.

Provisioned per contract

Shared queue

Your character in video

Standard plans animate any still. A private model keeps the character on-model.

Animated from your model

Generic i2v

Adult and furry content allowed

Billing

Neither auto-charges. We keep no card.

Invoiced per period

Self-serve, one payment

Request a custom plan

What needs to stay consistent, and roughly what you expect to generate. The volume figure sets the capacity we quote. We answer within one business day.

Frequently asked questions

Who owns the model you train for me?

You do. It stays scoped to your account and is never merged into the public checkpoints. Ask us to delete it and the training set goes with it.

Do my references or generations train anything else?

No. Nothing you send and nothing your private model produces is exported to the dataset.

Is adult furry content actually allowed?

Yes. Same limits as the public site: everyone depicted is adult and fictional, and the prompt filter still applies.

How long before it is running?

About a week for a first Creator Custom model. Two to three for a dedicated API.

What exactly is a clip?

720p, 5 or 10 seconds, with audio, animated from a still. It is image-to-video, so the still comes from your model first, which is what keeps the character on-model.

How does billing work?

Setup once, then monthly, invoiced per period. We hold no card on file, so nothing is ever charged automatically.

What if I only need a bigger quota?

Then skip this. Standard plans already go to unlimited images.