Anymal OSDraftv0.1

An AI operating system for the world's food systems

Our mission is to make the world's food systems legible to AI and embodied robotics. Food systems are networks of biological assets moving through production, logistics, compliance, finance, and consumption. Anymal OS gives every asset a shared, source-grounded representation, a coordinated way to act on it, and a trust rail any party can verify.

01

The scale of the problem

A $10T economy that runs mostly on paper. Numbers, not adjectives.

The world's food system is the biggest economy on the planet that humans don't directly sense. Most of its data still lives on paper, in cash, and in the heads of the people doing the work. The cost of that opacity shows up everywhere: in disease losses, fraud, foodborne illness, trapped capital, and a third of the world's farmers who can't prove what they have.

$10T+
Hidden cost of the global food system, annually
source · EAT-Lancet · Oxford, 2024
84B
Land animals slaughtered each year
source · FAOSTAT, 2024
570M
Farms worldwide · 475M of them under 2 hectares
source · FAO
Livestock

84 billion land animals enter the global slaughter system every year. Identity, ownership, health, and movement records still live on paper, ear tags, and brand books across most of the world — including in developed economies. When disease moves, the cost compounds before anyone sees it.

374M tglobal meat produced in 2024FAO
$111BChina alone lost to a single disease (ASF)WOAH
Mostsmall-scale farmers keep no electronic recordsFAO traceability review
Crops + Land

Primary agricultural production is a ~$5T direct output. $2.22T in food crossed borders in 2025 alone. And 475 million of the farms feeding the world operate on under 2 hectares — the people producing a third of the global food supply can't prove what they have to a lender, a buyer, or a regulator.

$5Tgross value of primary agricultural production / yrFAO
$2.22Tglobal food import bill, 2025FAO Food Outlook
0.8%share of global climate finance reaching small farmsIFAD
What the opacity costs · every year
$30–50B / yr
Food fraud, undetected because there is no shared identity
source · PwC · FDA
600M / yr
People sickened by what they ate · 420,000 died
source · WHO FERG
$170B / yr
Unmet smallholder credit demand (of $240B requested)
source · ISF Advisors · IFC
$300B → $1T
Sub-Saharan Africa smallholder output, today vs 2030
source · IFC
Notary. Every number above carries a named authority · FAO, WHO, PwC, IFC. The Notary signs and anchors any cited fact to a verifiable record, so a buyer, an auditor, or a counterparty can verify the source without trusting our database.Try the Notary →
The data-input failure. A majority of small-scale farmers in developing economies keep no electronic records at all (FAO). Those who do still use ear tags, brand books, and handwriting. Even in developed economies, ranch-level cattle records and water-rights diligence run on scanned PDFs and county clerks. This is why AI agents have not yet been able to reason across food systems · there has been nothing legible to reason over. Anymal OS exists to change that.
02

Why now

Three shifts converge. The window opens once.

Shift 1
AI agents reason over physical-world systems

Tool-use, structured output, and multi-agent orchestration crossed the reliability threshold. Agents can now run diligence that used to need a human walking a field.

Shift 2
Per-animal sensors are scaling

Biological assets are becoming continuously observable for the first time. Cattle, soil, water, and crops now emit evidence an agent can read.

Shift 3
Verifiable settlement matured

Fast, cheap on-chain attestation is production-ready. Agents can pay each other, sign their work, and produce claims any third party can verify without trusting our database.

03

Food systems are biological assets

Strip away the verticals and one structure remains: biological assets, constrained by the physical world, moving through a supply chain. That is the surface area.

Livestockactive
cattle, sheep, goats, bison, poultry
Cropsfuture
grains, produce, forage
Geneticsfuture
breeding lines, seed, semen
Soil biologyfuture
microbial carbon, fertility
Carbon capacityfuture
sequestration, insetting
Perishablesfuture
biologically constrained inventory

Livestock is live today; the rest are where this expands. Each asset is anchored by an Anymal ID, its persistent identity and lifelong record.

04

There are layers to this

Five layers, top to bottom. Four are running in production. The fifth is what makes the others portable, payable, and independently verifiable.

Anymals act. Sages know. Notaries attest. The Ontology binds them. Sources feed them.

Proof
Attestation Network
Building

A trust rail for agent-to-agent verification. Anymals do the diligence. The Proof Layer is how they get paid for it, sign their work, and produce claims any party can verify without trusting our database.

Action
Operations Layer · Anymals
Live

Agents that operate organizations across food systems: supply, compliance, logistics, insurance, banking, price discovery, and settlement.

Knowledge
Earth Layer · Sages
Live

Domain-specialist agents that watch over a place and know it. Water, soil, land, carbon, cattle. Each Sage observes its domain continuously and emits typed evidence with source, vintage, and confidence on every fact.

Substrate
The Evidence Ontology
Live

One parcel-anchored, source-cited representation of every biological asset that all agents reason over.

Foundation
Authoritative sources
Live

Government, open, licensed, satellite, and sensor data. Never the open web.

Show all sources ▾
Parcels & boundaries
  • US Census TIGER· county geographies
  • TxGIO StratMap· Texas statewide cadastre
  • TN Comptroller GeoViewer· Tennessee statewide
  • NSW DCDB· New South Wales cadastre
  • Per-county ArcGIS cadastres· 15 US states + NSW
Soil
  • USDA SSURGO· soil survey
  • USDA SoilWeb WCS· mukey raster tiles
  • USDA EDIT· ecological site descriptions
  • USDA cocropyld· soil crop yields
Land use & cover
  • USDA CDL· Cropland Data Layer
  • USDA NASS· national ag statistics
  • RAP· Rangeland Analysis Platform
Water · physical
  • USGS NHD· National Hydrography Dataset
  • USGS NWIS· streamflow + groundwater
Water rights · 10 US jurisdictions
  • IDWR· Idaho
  • CDSS· Colorado
  • DWRi· Utah
  • DNRC· Montana
  • OWRD· Oregon
  • WA Ecology· Washington
  • TCEQ· Texas
  • SEO· Wyoming
  • OSE· New Mexico
  • ADWR· Arizona
Animal sensors
  • Per-animal sensors· location + biometrics
  • Calving telemetry· dam-tag activity
  • Receiving-scale telemetry· packer intake
  • Cold-chain reefer logs· freight integrity
05

Live today

What is actually built and running. Source-cited, parcel-anchored, queryable.

38.46M
parcels in the spine
15
US states + NSW
14
US water-rights jurisdictions
5
domain Sages shipping
4
reference applications

Every fact in the system ships as a SpineFact: a typed envelope carrying value, source, vintage, and confidence. No claim without provenance.

06

What you can do today

Four reference applications running on the substrate. Each is a distinct lens; all read the same evidence.

WATER SAGE
Water-rights diligence

Source-cited water-rights diligence across 14 US states. Decreed use, priority ladder, supported parcels, owner alignment, regulatory overlays, NHD source geometry, live streamflow and groundwater conditions. Memo-ready in seconds.

11 SAGES + ANYMALS
Ranch-partner insetting

Locate a processing facility, set a haul radius, watch 11 agents score and rank ranches for joint Scope-3 + margin partnership. Sages gather, Anymals decide. Every claim cites the substrate it came from.

CATTLE SAGE
Procurement forecasting

Carcass yield, composition forecast, and procurement bias from a cattle inventory. A buyer sees the whole diligence chain narrate on screen as the Sages work.

CARBON SAGE
Carbon-headroom shed

County and parcel-level ranking on sequestration headroom. Soil-organic-carbon stock, forage capacity, ag fraction, all cited to USDA SSURGO, NASS, and RAP.

07

What's next

The Proof Layer · agent-to-agent verification on Solana. Sages and Anymals already do the work. The Notary is how it becomes paid, signed, and independently checkable.

BuildingAgent-to-agent verification on Solana. Under active development.

The Sages already do the grunt work of verification. The Proof Layer is how that work becomes paid, signed, and independently checkable. Notaries orchestrate the payment and the attestation. The four lower layers stay private. Only the commitment crosses the boundary.

HOW IT WORKS
Lender agent  ·  requests cattle-collateral diligence
                       │
                       ▼   Solana   ·   payment
                       │
                Anymal Notary
                       │   (orchestrates · signs · anchors)
                       │
        ┌──────────────┼──────────────┐
        ▼              ▼              ▼
    Cattle Sage    IDx Registry   Carbon Sage
   (life, breed,   (ownership,    (forage,
    location)       holdings)      headroom)
        │              │              │
        └──────────────┼──────────────┘
                       │
                       ▼   signed attestation
                       │
                       ▼   anchored on Solana
                       │
        Lender  ·  insurer  ·  buyer  ·  regulator
                  verify independently

Raw evidence stays off-chain: animal identity, owner identity, GPS history, sensor logs, ranch records. On-chain stores only the compact proof artifacts: payment tx, evidence root, proof hash, attestation hash, issuer wallet, request id, status, timestamp.

What the Proof Layer unlocks

Lending
TODAY
Manual inspection, PDFs, rancher statements, trust in intermediaries.
WITH THE PROOF LAYER
Cattle as legible collateral. A lender agent pays for diligence, gets a signed attestation, anchors the proof, and underwrites against verified livestock.
Insurance
TODAY
Pricing from estimates, stale records, and self-reported herd inventories.
WITH THE PROOF LAYER
Pricing from real evidence. Continuous observability + signed attestations let insurers price risk against measured stocking, biosecurity, and forage condition.
Carbon revenue
TODAY
Modeled outcomes, manual MRV, multi-month audits.
WITH THE PROOF LAYER
Measured outcomes. Carbon claims anchored to per-animal sensor histories and substrate-cited Anymals. Buyers verify without re-running the work.
Traceability
TODAY
Claims travel as marketing copy and certifications you have to trust.
WITH THE PROOF LAYER
Gate-to-plate as a chain of attestations. Each handoff signs, anchors, and remains independently checkable years later.
Audit
TODAY
Site visits, document review, point-in-time inspections.
WITH THE PROOF LAYER
Cryptographic audit. A regulator verifies a claim with a wallet check, not a flight.
Open marketplace
TODAY
No way for one company's agent to pay another for verification.
WITH THE PROOF LAYER
Sages and Notaries become payable services. Other agents hire ours and receive signed attestations. Reputation accrues to issuer wallets. The marketplace is open.

Anymal OS becomes infrastructure for an agent economy on biological assets. Other companies' agents pay ours. Reputation accrues to wallets. The marketplace is open.

08

Built to scale globally

The pattern is universal, not the data. We don't need new world infrastructure. We adapt to what every jurisdiction already has.

Every jurisdiction has authoritative sources

Foundation Layer is government + sensor + satellite. That stack exists in every food-producing geography. We don't build new data. We adapt to what already exists.

Parcels and IDs are universal primitives

Land has boundaries everywhere. Animals get identifiers everywhere. The Evidence Ontology is a shape, not a country.

The Anymals replicate per geography

Soil, water, land, carbon, cattle. The pattern repeats. We proved it: Idaho first, then 14 US states, then NSW. New geographies onboard in weeks, not quarters.

The Proof Layer is chain-native

Attestations don't need jurisdictional install. Solana runs everywhere. The trust rail crosses borders by construction.

The wedge

We start with animal infrastructure, the livestock wedge, and expand into broader biological-asset coordination across protein, crops, compliance, logistics, finance, and traceability.

How it all works, the evidence ontology →Watch 11 Anymals run insetting →Explore the parcel spine →
Source-backed context only. Every output keeps facts (source-observed and cited) separate from interpretation, and Anymals draft actions for human approval. Not legal, financial, or trading advice.