Veldarium
Systems

Four public systems. Four hard operating domains.

Each system tests the same company thesis from a different edge: records, approvals, exceptions, memory, and human-controlled AI execution.

Animal placement OSPublic build track

WhiskerMatch

Animal placement OS for intake, fit review, foster capacity, adopter handoff, follow-up, and outcome memory.

Market
Animal placement / shelter and rescue
Broken workflow
Notes scatter. Capacity changes. Adopter fit is judged late. Follow-up disappears. Returns become memory nobody can query.
First controlled workflow
Placement queue and dossier for one shelter or rescue workflow.
Who it is for
Shelters, rescues, fosters, adopters, placement teams
What validates it next
Validate one real placement workflow from intake through follow-up with a shelter or rescue design partner.
Open dossierSee the architecture
Regulated agriculture OSPublic build track

AcreFrame

Regulated agriculture OS starting with licensed cannabis operations where batch history, SOP drift, labor, testing, yield, and traceability matter.

Market
Regulated agriculture / licensed cannabis operations
Broken workflow
Batch records drift. SOPs mutate. Testing creates pressure. Labor notes fragment. Compliance review happens too late.
First controlled workflow
Batch control packet for one licensed cultivation or production workflow.
Who it is for
Licensed operators, cultivation teams, compliance leads, QA teams
What validates it next
Prove one batch-control workflow with synthetic packets, then validate field structure against a licensed operator's real process.
Open dossierSee the architecture
Food distribution OSPublic build track

Fresh Margin Systems

Food distribution and fulfillment OS for supplier-to-shelf, kitchen, truck, invoice, and customer outcome.

Market
Food distribution / margin-sensitive purchasing
Broken workflow
Supplier variance hides in invoices. Substitutions leak margin. Spoilage gets normalized. Credits are missed. Buyers decide without a clean record.
First controlled workflow
Supplier variance and purchasing exception board for one buyer or category.
Who it is for
Distributors, food operators, procurement teams, category buyers
What validates it next
Validate one supplier-variance workflow with real purchasing logic, synthetic data first, then design-partner records under agreement.
Open dossierSee the architecture
Shipyard & heavy industrial execution OSEarly architecture

STBD.ai

Shipyard and heavy industrial execution OS for steel, labor, assets, drawings, work packages, inspections, blockers, and yard execution.

Market
Shipyard / heavy industrial execution
Broken workflow
Materials stall crews. Drawings change. Inspection hold-points sit unresolved. Work packages drift. The schedule lies until the yard feels it.
First controlled workflow
Blocked work-package register for one yard area or production cell.
Who it is for
Yards, fabrication teams, maintenance crews, inspectors, engineers
What validates it next
Scope one blocked-work workflow with domain operators before claiming operational readiness.
Open dossierSee the architecture
At a glance

How the four systems differ.

Domain
WhiskerMatch — Animal placement / shelter and rescue operations
AcreFrame — Regulated agriculture / licensed cannabis operations
Fresh Margin Systems — Food distribution / fulfillment / margin-sensitive purchasing
STBD.ai — Shipyard / heavy industrial execution
Messy inputs
WhiskerMatch — Animal histories, adopter applications, foster observations, behavior notes, home context, transport updates, and follow-up messages.
AcreFrame — Task notes, batch context, SOP fragments, production schedules, environmental readings, test requirements, labor notes, and incident reports.
Fresh Margin Systems — Supplier quotes, invoices, order history, receiving notes, substitutions, credit disputes, category messages, and customer issue records.
STBD.ai — Work packages, drawings, change orders, inspection notes, crew assignments, material updates, supplier messages, and equipment status.
Core objects
WhiskerMatch — Animal profile, placement dossier, fit review, foster capacity state, adopter readiness state, return-risk flag, follow-up record.
AcreFrame — Batch plan, SOP state, drift alert, compliance review packet, QA note, yield/margin memory.
Fresh Margin Systems — Quote/invoice intake, supplier variance, purchasing exception, fulfillment issue, margin memory.
STBD.ai — Work package, material blocker, crew/asset state, inspection gate, production memory.
First artifact
WhiskerMatch — Placement Dossier
AcreFrame — Batch Control Packet
Fresh Margin Systems — Supplier Exception Report
STBD.ai — Yard Blocked-Work Brief
Exception type
WhiskerMatch — Return risk / fit mismatch
AcreFrame — SOP drift / batch gap
Fresh Margin Systems — Supplier variance / margin leak
STBD.ai — Material shortage / inspection hold
Why horizontal AI fails here
WhiskerMatch — A prompt cannot own placement history, return risk, foster capacity state, follow-up accountability, or the operating loop between shelter, rescue, foster, and adopter.
AcreFrame — Generic AI has no facility-specific batch history, SOP state, environmental signal context, compliance review boundary, or traceable audit record.
Fresh Margin Systems — A chat interface cannot see vendor history, receiving state, SKU economics, margin memory, or the exception loop between buyer, supplier, and fulfillment.
STBD.ai — Horizontal tools do not understand work packages, inspection hold-points, yard state, material dependencies, crew sequencing, or release gates.
Current status
WhiskerMatch — Public system surface active. Real placement pilots require shelter and rescue partner validation.
AcreFrame — Public system surface active. Compliance-aware workflow structure needs validation with licensed operators.
Fresh Margin Systems — Public system surface active. Pilot-ready workflow scoping with purchasing-heavy food operations.
STBD.ai — Early architecture. Public language remains bounded while shipyard workflow scope is validated.

The portfolio is focused.

WhiskerMatch, AcreFrame, Fresh Margin Systems, and STBD.ai are the public systems. The next step is a real workflow to test against one of them.

Start a founder conversationView synthetic preview