For investors
Nobody sells a complete, accurate set of animals. We are building one.
Games, films and museums that need accurate animals build them in-house, a few hundred at most, and never license them out. We are building animals from real scans and measurements, with every source on record, on a pipeline that makes the next species the same way.
The market we see
Accurate animals are made in-house, one studio at a time.
From our own market research (October–November 2026). Each figure links to its source. We are still checking some of them.
The biggest are about 200 species
Red Dead Redemption 2 has about 200 species (Rockstar); Planet Zoo has 76 plus 134 in add-ons (Wikipedia). Each was built by a team over years, and neither is licensed out. Documentary studios build their creatures per show.
AI floods uploads, not sales
On CGTrader, AI-made models were about 1 in 6 uploads but about $1 in every $90 of revenue (CGTrader, 2026). TurboSquid bans AI-made models (TurboSquid). 52% of game developers now call generative AI harmful (GDC 2026).
Animals that bend and move
AI generators make shells, not animals that walk. Commercial AI animation gives a four-legged animal one walk cycle or 16 generic presets. Our fox has 105 clips generated on its own measured skeleton. We found nobody selling measured, cited, rigged and animated animals.
A crowded shop window
Fab, Epic's store, passed 420,000 listings in 2025 (Epic). Realistic rigged animals there and elsewhere mostly sell for $100–350 each, by our survey of listings. Selling on accuracy means showing the proof on every listing.
How it makes money
Models first. The rest comes later.
We plan to earn from animal models first, then from library and science licences once a species is signed off. The game and the mod market are long builds that will not earn for years. Every line uses the same animals, plants and land, so work on one feeds the others. Prices, splits and stores are not set yet.
Animal models
Rigged, animated animals with Blender source files and Unreal-ready exports, each with its source sheet, sold through the big stores.
Library and science licences
The whole library for a studio, and licences for museums, domes, vet schools and research. Fewer, larger and slower deals that need a finished, signed-off species to show.
The game
A world simulation from the end of the Ice Age to today, sold on its accuracy. Store not chosen yet.
The mod market
Creators build and sell worlds with their own rules on our base, and we take a share of each sale. The split is not set.
What we think stays ours
Generic animals will get cheap. Evidence won't.
We expect generic rigged animals to be close to free within a few years. What a generator cannot copy is what each of our models carries.
- Measurements and scans with a pedigree. Every input, licence, build step and check score is recorded, and the sources page is public.
- Partner data. Scans, footage and notes from museums, zoos and researchers that are not online anywhere, credited in every model.
- One skeleton standard per body plan, so motion and tools carry from one species to the next.
- Motion from measured gaits, at real speeds and strides, which no clean dataset offers for animals.
- A pipeline of rules, not hand-work, so species two to a thousand are built the same way.
Where it stands
Two months in, one animal from sale.
The build started in September 2026. Numbers come from the project's own records and update each time the site is built.
- Red fox: built from one real scan, rigged, furred, with 105 motion clips. Paws and some leg poses are being fixed before it can sell.
- Common raven: started from an open scan; its feathers passed their first test.
- Moose: built from an open scan of a bull; its walk and trot are in testing.
- About 47,700 animal species with a rough draft from open trait data (measurements and rule-built shapes, not models), and 365,813 plant species in the tables.
- The game: a simulation of the whole planet from the end of the Ice Age to today, where you can grow from one hunter's camp to a city or a realm. 13 of 14 steps to its first playable version have passed.
- The full plan is on the build progress page.
Risks, as we see them
What could go wrong.
We would rather you read these here than find them later.
- Nothing is on sale yet. The fox still has to pass its checks.
- Unproven premium. We have not yet shown that buyers pay more for verified accuracy. A small store test of the first species is the cheapest way to find out.
- Data supply. Open scans cover about one species in 300. Growth past that depends on partners, and no partner terms are agreed yet.
- Discovery. The stores are crowded and flooded with AI assets, so being found is hard.
- Slow big deals. Science and library licences are large but slow, and need expert sign-off we do not have yet.
Talking to us
No round is set yet.
We have not set an amount, terms or a timeline. If you would like to follow the project or talk when we do, register interest and choose "Investor".
Where money would go first, as we see it now
- Scans at scale: capture trips and data deals with museums, zoos and parks.
- Expert sign-off: zoologists and vets checking each species before it sells.
- Compute: more GPU time, so species are built in parallel rather than one at a time.