Nerio × Minerva, officially launched
Governed intelligence: the trust layer for enterprise AI, with a read side and a write side.
AssetShop EOS is now a complete enterprise operating system. Nerio sees: it reads every system a company runs, changes nothing, proves where the records disagree, and serves the company's AI only what is evidenced, with every figure labeled. Minerva acts: it executes only what a person has authorized, records the reversal before the write, and keeps every action on a chain an auditor can verify without us. Honos decides who may act, on what evidence, and asks a second person above the threshold. Every decision reconstructable. Every action reversible.
No write method exists in its connector contract. Its promise on every page of this site: Nerio never writes to your systems of record.
Five rules, each a test: no execution without an approved request; every write reversible; dual control above the threshold; idempotent on the request; hash-chained per organization. Ledger execution today; each ERP adapter is listed with its status the day it ships.
The Enterprise Operating System holds both intelligences, Honos and the record. Two delivery models: Hosted (the control plane, named sub-processors, a DPA path) and Sealed (the air-gapped artifact, no runtime calls).
Proprietary intellectual property
What we own, what we publish on purpose, and what cannot be copied by tomorrow.
Held. The reconciliation engine's rules and claim ranking; the label contract of the context service; the evaluation battery; the Minerva execution model. Published on purpose. The Verify Standard and the Proof Object schema, under CC BY 4.0, because a standard only creates a moat if others can adopt it. Marks. ASSETSHOP, then the composites ASSETSHOP NERIO and ASSETSHOP MINERVA. Candidates for filing, subject to counsel. Compensating-action-before-commit on a per-organization hash chain; a labeled context service whose principal schema has no execute authority.
The moat is earned, never manufactured: the per-organization record that accumulates from the first engagement; policies, versions and clauses tuned by the customer; read-only adapters that need no security review to add; execution that records its reversal first; and honesty as a control, where the page says what does not exist yet. Exported evidence verifies after a customer stops paying us, which is a deliberate constraint on our own leverage.
In commercial production, sold by engagement
An assessment, a pilot refunded against agreed targets, then an annual platform agreement.
EOS is in commercial production and is sold by enterprise engagement only. Nothing is bought from a page. Value is priced per governed principal, per connected system and per governed decision, never per human seat, so expansion follows connectors, principals, decision types and executed actions. The $15,000 assessment is credited in full against the $50,000 pilot; the $35,000 balance is fully refundable if the mutually agreed target is not achieved by day 90.
The market we serve. Resilience is the operating condition: every organization must now prove its AI can be trusted with consequential decisions, to auditors, regulators, boards and owners. We expect demand to compound, because each system connected raises the value of the next, each decision type governed opens the one beside it, and each executed action makes the next safe to automate. That expectation is stated as an expectation.
Read the public record for the state of every stage, then the diligence register, the trust center and the covenant. Forward-looking statements are expectations. Nothing on this page is an offer of securities.
Investors · ungated · August 21, 2026
The build is stated. The market is yours to model.
This page assigns no pre-money. It states what exists today, what opens next, and hands you the market model, sliders, your assumptions, your number. Every build figure below was reproduced by executing the shipped artifact rather than asserted.
Where the build stands today
Stated plainly, and reproducible: the numbered items below are what a visitor can execute or open right now.
- The agreement layer. EOS runs today as a full platform demonstration on a synthetic tenant: 22 conformance suites, 403 cases, 0 failures, executed cold and offline by an external party. Nerio is read-only by construction, and the write-path battery proves it has no write method; Minerva writes only on a named approval.
- Portunus. Six live synthetic assets reconciled across thirteen dimensions, every figure tracing to evidence, the consumer-verification digest recomputed from source and matched to its anchor.
- The Verify Standard. Ratified as an adoptable open specification; the estate conforms at V3 and the diligence harness is published unauthenticated, so verification needs no account and no vendor in the loop.
- Capacity Redeployment. The newest capability: freed hours and KPI state in, a ranked, arithmetic-visible plan out, adoption gated behind a named authority. Run it.
- The commercial machine. A live estate API with card checkout armed end to end in test mode, entitlement webhooks, and a payments ledger; the consumer estate carries the first paying surface.
- The estates. Ninety-six enterprise pages and seventy-seven consumer pages, every page under an automated gate battery for claims, lexicon, contrast, security, and portability, with product films covering all nine doors that carry app code.
Opening next
- EOS Maverick: Access as Proof. The mission-assurance front door, live and public. Where access itself must be provable, Maverick is the conversation opener. The page is the pitch.
- Live-mode commerce. The test-mode machine cuts over to live keys in one planned motion; the machinery is already deployed.
- The pilot program. The capital plan below points at paid pilots on production estates, measured under the published standard; the public record prices the next round.
- Independent security assurance. Funded in this plan; the observation window is announced when it is set, not before.
What was independently verified
An external party executed the shipped artifact cold, offline, with no network and no vendor in the loop. These are reproductions, not assertions. They verify the machine, not the outcome; the outcome column opens at stage four of the public record, and a single undeniable customer result will outrank every row below.
| Check | Result | Method |
|---|---|---|
| Conformance battery | 22 suites, 403 cases, 0 failures | Executed cold, offline |
| Full estate battery | 14,914 of a stated 14,914 assertions | Vendor instrument, independent run. Gap published |
| Release receipt | Re-derived exactly | 860 sealed entries, VERIFY clean |
| Tenant isolation | 45/45 | Executed |
| Write path | 9/9, no write method exists | Executed |
| Connectors | 8 conformed, 9 declared | 66 adapter cases |
| Codebase | 8,477 source, 5,184 test lines | Counted |
Verified in the build since August 13, 2026, not yet independently
Self-verification by the build, dated August 15, 2026, each with its instrument named. It is not the external run above and is not counted as one.
| Check | Result | Method |
|---|---|---|
| Trunk release-integrity gate | PASS, 0 P0, 0 P1 open, 2 not testable (a real vendor cost; a live origin), 2 lab-measured | eos-tools.py gate over the shipped estate |
| Trunk conformance suite (backend) | 34 files, 823 assertions, 0 failures | re-run offline in the build container |
| Estate battery | eleven instruments green over 1,010 sealed files | verify-estate.sh, cold from the pristine master |
| Tenancy kernel | 91 of 91 checks (two tenants, eight identities, cross-tenant paths, lifecycle) | local battery, stand-in SDKs, real HTTP; not deployed |
| Estate API | 88 routes PASS, 4 blocked on external dependencies, 3 of 3 webhooks idempotent | local runtime harness; not deployed |
| Immersive engine | selects the WebGPU renderer with WebGL2 fallback; zero errors in 29 realms | headless Chromium; devices pending |
| Enterprise platform | DEMO on a synthetic tenant; commercial capability verified 0 of 18; backend source for about 12 of 18 in the trunk, unverified at runtime | MASTER_COMPLETION_MODEL.md |
The market, sized by assumptions you control
This model is yours. The company assigns no valuation on this page; the defaults are assumptions stated so they can be argued with, and the next round prices off the public record, not off this instrument.
The investor room
Per-product detail sits behind a passcode, because it is written for people already in a conversation rather than for search engines. If you have the code, use it. If you do not, ask; the form takes a minute and a person reads every one.
Request access
Tell us who you are and what you are evaluating. AssetShop reviews each request and releases a passcode where it fits.
Use of capital
The allocation is the plan's structure and holds at every round size; the dollars follow the round. The market assumptions above describe the outcome, not what the next 24 months cost, so they do not move this table. Line descriptions are written for the $6M plan.
| Line | Amount | Share | What it buys |
|---|---|---|---|
| Commercial proof | $2.1M | 35% | Three to five paid pilots on production estates. Converts the starting line above into measured customer outcomes. The only line that changes the valuation. |
| Engineering | $1.5M | 25% | Two senior engineers. Promote 9 scaffolds to conformed in production engagements, deploy, ship billing and entitlements. |
| Security and assurance | $900k | 15% | SOC 2 [not claimed] Type II observation window and report, penetration testing, source escrow [not claimed], patent counsel. |
| Go to market | $900k | 15% | One operating-partner-facing AE, one solutions engineer, category content, the Verify Standard conformance register. |
| Legal and corporate | $300k | 5% | C-Corp conversion, MSA/DPA/SLA to signable instruments, trademark, cap table. |
| Operating reserve | $300k | 5% | Three-month buffer beyond the 24-month plan. |
The milestone this capital is judged against: at month 12, three signed pilots with measured findings and a completed SOC 2 [not claimed] observation window.
Assumptions, listed so they can be argued with
Every item below is an assumption, not a measurement.
| Assumption | Status |
|---|---|
| Enterprise and sponsor buyers pay for evidence and reconciliation as a category | Untested. The central bet. |
| A value-share model is acceptable to procurement and to fund finance | Untested. |
| 9 scaffolds promote to conformed at the cost of the 6 derived adapters | Reasonable; validates in the first engagement. |
| A private equity motion reaches 3 to 5 pilots in 12 months | The strongest available motion, still unproven. |
| A real tenant behaves as the mock transport does | The largest technical unknown. |
| Category language holds against an incumbent adding reconciliation | The largest competitive risk. |
| Portfolio products acquire users without paid distribution | The weakest assumption in the document. |
The starting line, priced in
EOS establishes decision-grade agreement across conflicting systems, with provenance and context, before downstream intelligence acts. The raise exists to buy commercial proof, and the plan is priced against the true starting position rather than an implied one. Stated once, plainly, at every level.
- The enterprise platform runs today as a demo on a synthetic tenant; the tenancy kernel and estate API are verified locally. The web estates are static software, production-ready as such.
- Production ERP engagement: the largest technical unknown, listed as such in the assumptions register.
- Customer proof: the pilot program’s deliverable, measured under the published standard.
- Independent security assurance: funded in this raise, not yet begun. Timing stated when an observation window is set.
- MSA, DPA and SLA: drafted as outlines for counsel, executed with the first pilot.
- The entity is an LLC. Delaware C-Corp conversion is a condition precedent to any priced round.
- The full battery reproduced 14,914 assertions across 146 suites against a stated 14,914 across 146. The 185-assertion gap is published rather than rounded away.
The questions a serious diligence asks
An independent analysis of this company produced thirty-six questions an investor would ask before a dollar moves. They are now answered on a public page, or marked open where nothing exists yet, or marked data room where a public summary could flatter. Four of the answers belong here too.
- The buyer. The CFO for a multi-ERP enterprise, the operating partner for a fund; Internal Audit champions. A working hypothesis, labeled as one.
- SNX. Closed-loop prepaid service credit on the consumer estate. No token, no contract address, no supply, no market, and none planned. The enterprise product has no SNX surface anywhere in it.
- The moat. Not the hashes, which any competent team can reproduce. Independence from every system of record, afternoon deployment, and a reconciliation corpus that compounds with every customer engagement. The ceiling is not reconciliation: if the architecture holds, EOS is the contextual substrate between enterprise reality and machine intelligence, and every integration and interaction compounds it; reconciliation is the wedge that earns the right.
- The focus rule. Until three signed pilots carry measured findings, EOS commercial proof outranks every other workstream, and nothing new ships on the consumer side.
The diligence register: all thirty-six, answered or marked open.
The next step
Ask for the command. The harness runs the battery cold on your own machine, offline, with no account and no vendor in the loop. Run it before the first call, then tell us what your model says. The seven-stage public record shows which stages are complete and which come next; the next round prices off that page. Per-product detail sits in the investor room below.
Request the diligence package