The Dog House

Serious dog mixer, adoption-fit engine, and visual breed lab.

Elite mixer ready
Male breed
+
Female breed
Checking assets/dog-preview.gltf
The Dog House real image preview
Visual preview is a breed-informed approximation. Mixed-breed appearance varies by individual genetics.

Compatibility
Weight
Height
Lifespan
Challenge

Research Hub

Research sits behind the mixer. The first screen stays visual and fun, but the logic underneath is built around dog retention, realistic ownership, DNA limits, shelter pressure, and behavior science.

Shelter pressure is a systems problem

Recent national reporting still shows millions of dogs and cats entering shelters each year. Shelter Animals Count estimated 2.8 million dog/cat intakes in the first half of 2025, while ASPCA reported 4.2 million shelter adoptions and about 597,000 euthanasias in 2025.

  • Match people to dogs they can keep, not just dogs they like.
  • Surface housing, budget, time, and behavior risks before adoption.
  • Route struggling owners to help before surrender becomes the only option.

Why dogs get surrendered

AAHA/HASS owner-surrender work points to repeat pressure points: housing challenges, too many pets, no time or overwhelm, financial constraints, and short-term life crises.

  • The app should ask about lease restrictions, pet rent, backup housing, and moving risk.
  • Cost warnings should include routine care, emergency bills, grooming, training, and boarding.
  • Keep My Dog mode should be visible, fast, and nonjudgmental.

Behavior is genetics plus learning history

AAHA, Merck Veterinary Manual, and WSAVA all frame behavior as a mix of inherited tendencies, socialization, health, stress, trauma, training, and daily environment. Breed helps with priors; it cannot read the individual dog.

  • Flag fear, anxiety, reactivity, impulsivity, and handling sensitivity as real owner-fit issues.
  • Separate normal workload from danger signs that need a veterinarian or qualified behavior professional.
  • Use first-90-days check-ins because shelter behavior can change after decompression.

DNA improves confidence, not certainty

UC Davis VGL, Cornell, and current dog-genetics research support DNA as a useful confidence tool for ancestry, coat traits, drug sensitivity, and some inherited risks. It is not a prophecy for final looks, behavior, or health.

  • Keep image previews strict photoreal and label them as breed-informed estimates.
  • Use DNA language for probability bands, not fake certainty.
  • Recommend vet review for health concerns and parent-breed screening questions.

Breed labels are weak evidence

A PLOS ONE study found shelter breed labels can affect perception and adoptability, especially for pit-bull-type dogs, while visual breed identification often disagrees with DNA results.

  • Show breed mix results, but match adoption by individual needs and household capacity.
  • Ask for foster notes, behavior observations, medical records, and meet-and-greet results.
  • Avoid making breed labels the only reason a dog is recommended or rejected.

First 90 days: the story keeps changing

The adoption report is not finished when the dog comes home. The app should track decompression, accidents, separation issues, resource guarding, kid/pet stress, medical concerns, and owner regret before return pressure builds.

  • Day 3: eating, potty, sleep, hiding, barking, and stress check.
  • Day 7: kids, pets, alone time, landlord, destruction, and regret screen.
  • Day 14/30/60/90: escalate to trainer, vet, housing, food, or foster resources.

Content Rules

Image realism rule

No animation style. Every preview should be photoreal, breed-informed, and visibly different when male/female breed order changes.

Story rule

Each story should be funny enough to remember, but it must teach a real management point: energy, handling, cost, training, housing, or health.

Matching rule

Dog requirements come before breed stereotypes. The right answer may be adult, foster-verified, low-drive, senior, or "not now without support."

Safety rule

Bite risk, severe aggression, medical symptoms, poisoning, collapse, repeated vomiting, or major anxiety routes to a veterinarian or qualified professional.

Dog-Fit Quiz Modes

The quiz supports the mixer. The Dog House can stay playful for a quick share or get serious when someone is making a real adoption decision.

Pick a lane. The playful one stays lighter and kid-safe. The serious one asks about housing, schedule, budget, training, family, pets, patience, and support so the app can give actual dog-fit results instead of just vibes.

Playful mode keeps the hook fun while still nudging users away from bad dog matches.
10 multiple-choice questions Dog types before breeds Support routing, not judgment

Best Dog Types

Breed Possibilities

Keep My Dog Mode

I might have to give up my dog. This flow routes the owner to food, vet, training, housing, crisis foster, safe-haven, or supported rehoming options before shelter intake becomes the only option.

First 90 Days Coach

The match is not finished at adoption. This is the return-prevention schedule: decompression, behavior checks, cost checks, safety routing, and no-shame help before regret turns into surrender.

Emergency Routing

Shelter Pilot Dashboard

A staff-side proof of concept: build an individual dog profile, compare it to an adopter readiness profile, and get conversation prompts instead of a cold pass/fail decision.

Product Strategy

Working master version: The Dog House is a serious breed mixer first, with dog-fit and shelter support features around it. Cheddar becomes a tribute feature, not the product name.

Product truth

The app is not really about โ€œwhat dog are you.โ€ That is the door. The real product is a smarter adoption filter that protects the dog, the owner, and the shelter from bad matches.

Partner-safe shelter framing

Help no-kill shelters reduce overcrowding by matching dogs with owners who are actually equipped for them. This supports shelters; it does not attack them.

Hook line

Cute gets your attention. Compatibility keeps them home.

Feature Roles

The Dog House

The public brand and front door. It should feel like a real product for mixing breeds, reading reports, and checking owner fit.

Visual Mix Lab

The headline feature: parent breeds, hybrid previews, trait scoring, appearance estimates, and shareable reports.

Research Hub

The source-backed layer for shelter pressure, behavior science, DNA limits, breed-label bias, and first-90-days retention logic.

Dog-Fit Quiz

A supporting lane that turns lifestyle answers into readiness, risk, and breed-type guidance.

Keep My Dog

Retention support for owners under pressure before surrender becomes the only path.

Shelter Pilot

A partner-facing proof of concept for matching individual dogs with realistic adopter profiles.

Cheddar Tribute

The previous work stays honored as an Easter egg and memorial feature without taking over the product identity.

Build Order Locked

1. Mixer first

Lead with the two-breed experience, visual previews, and a report that feels worth sharing.

2. Breed reports

Keep the encyclopedia useful with temperament, health watchlists, owner fit, and source-backed language.

3. Research backbone

Keep shelter data, behavior science, DNA limits, breed-label bias, and source citations visible enough to earn trust.

4. Quiz structure

Use fun quiz language only where it helps, then collect real lifestyle data: family, housing, exercise, patience, training, budget, and stability.

5. Compatibility scoring

Score both sides: owner readiness, dog compatibility, risk, training difficulty, and cost fit.

6. Warning system

Show dogs to avoid, owner risks, and exact fixes before adoption.

7. Shelter pitch

Use the app to reduce returns, reduce overcrowding, and support better conversations without becoming class-biased gatekeeping.

Saved Mixes

Compare Breeds

Asset Credits

Current build

The realistic 3D lane is wired for a local Quaternius glTF at assets/dog-preview.gltf and a local copy of @google/model-viewer at assets/vendor/model-viewer.min.js. If the model or viewer is unavailable, the app uses a procedural SVG dog preview with breed-informed colors, ears, muzzle, build, and tail clues. The badge on the preview tells you which path is active.

Installed breed photo references

Breed cards use local Wikimedia Commons reference photos when a license-tracked image is available. Missing breeds keep the 3D/SVG visual lane until a cleanly licensed photo is added to assets/photos/. Full metadata is stored in assets/photos/_manifest.json. The five-image-per-breed reference gallery is stored at assets/breed-gallery/_manifest.json; the Elite image lab now uses those photos as background reference context while the foreground dog is a role-aware hybrid drawing from breed traits.

BreedLicenseSource
RottweilerCC BY-SA 3.0Wikimedia Commons
Siberian HuskyCC BY-SA 3.0Wikimedia Commons
German ShepherdCC BY-SA 2.5Wikimedia Commons
Golden RetrieverPublic domainWikimedia Commons
Labrador RetrieverCC BY-SA 2.0Wikimedia Commons
French BulldogCC BY-SA 4.0Wikimedia Commons
BulldogCC BY-SA 4.0Wikimedia Commons
BeagleCC BY-SA 3.0Wikimedia Commons
PoodleCC BY 2.0Wikimedia Commons
DachshundCC BY-SA 4.0Wikimedia Commons

Preferred 3D source

Primary source: Quaternius Ultimate Animated Animal Pack. It lists glTF support and a CC0/Public Domain license, making it the cleanest free route. The current preview uses the pack's Husky glTF as a dependable breed-informed 3D base.

StepAction
1Download the Quaternius pack from the source page.
2Find a dog-like glTF or GLB in the glTF/GLB folder.
3Rename the chosen entry file to dog-preview.gltf or update the viewer source.
4Place it at assets/dog-preview.gltf.
5Refresh this page. The 3D badge should switch from fallback to loaded.

CC-BY fallback policy

Sketchfab can be used only when a model is downloadable and has clear CC-BY or CC0 terms. For CC-BY, record the exact model title, creator, source URL, license, download date, and modifications in assets/ATTRIBUTION.md. Avoid paid, unclear, non-downloadable, editorial-only, or restrictive assets.

"dog-preview.gltf" by [Creator] is licensed under CC-BY 4.0. Downloaded from: [URL]. Modifications: renamed file, compressed textures.

Built-in image creator

The live no-spend image lane now renders every pair through the same sire/dam role-aware hybrid engine. Reference photos are used as soft background context, while the foreground dog is built from breed traits, coat genetics hints, body shape, ears, tail, life stage, and male/female role modifiers. Paid providers stay disabled until server-side keys, QA gating, and a spend cap are explicitly approved.

Prompt template: A highly detailed, photorealistic studio portrait of one crossbreed dog from a male [Breed A] sire and female [Breed B] dam, with [breed-weighting], realistic fur texture, natural eyes, breed-informed body shape, soft studio lighting, 50mm photography.

Copied