Genesis PreviewE1 admittedSemantic Snapshot implemented500 Atlas identitiesQuery Lab: liveRead-onlyNo arbitrary URLsHosted CI: startup blocked

Live read-only Genesis baseline plus protected E4.7 agent, ontology and model candidates. Public MCP, frontier-model execution and TNOE training are not deployed; no arbitrary URLs.

TWIRX Reference implementation and public service of the Typed Web Commons

Engineering and model roadmap

Grow utility, semantic diversity, and model intelligence without weakening proof.

The 90-day grant program builds five useful universes and a visual agent. A separate evidence-gated path takes TNOE from architecture scaffold to a powerful semantic-induction research system while keeping all model output candidate-only.

From verified infrastructure to useful semantic universes

The $30,000, 90-day program builds an executable ontology fabric, five coherent knowledge and capability universes, a visual Atlas Agent, semantic change streams, a small evaluated semantic model, and publisher and federation paths.

Packet totals remain subordinate to useful frames, source diversity, query coverage, proof completeness and genuine change evidence. E4.5 has already validated more than one million real source-derived packets; this program makes that state substantially more useful and diverse.

Milestone 1 — Ontology and Live Utility Alpha

$6,000 · days 1–30

  • Publish E4.5 and implement Ontology Fabric 0.1 with frame, mapping, semantic-diff and generated-validation contracts.
  • Build Opportunity and World State modules plus initial standards bridges.
  • Expand selected high-value World Bank indicators.
  • Reach at least 1.5 million real packets and 120,000 proof-linked frames.
  • Produce two useful materialized cross-origin views and two genuine source-derived delta demonstrations.
  • Ship a Visual Atlas Agent alpha with ten curated investigations, maps, timelines, comparisons and provenance views.
  • Establish a dedicated versioned TWIRX artifact-storage path and publish exact unit economics.

Milestone 2 — Five-Universe Semantic Web

$10,000 · days 31–60

  • Add Research and Evidence, Security and Standards, and Agent Economy universes.
  • Integrate at least eight primary structured source families.
  • Reach at least 2.5 million real packets and 250,000 semantic frames.
  • Publish five materialized universe views, four genuine source or semantic change streams, and twenty curated investigations.
  • Implement source-stated commercial capability semantics for APIs, operations, access, cloud offers, regions, units and prices.
  • Publish a reproducible browser, search, native-interface and TWIRX benchmark with exact scope.

Milestone 3 — Learning, Publisher, and Federation Alpha

$14,000 · days 61–90

  • Build WSIM Seed for surface classification, mapping and frame-slot candidates, ambiguity and abstention.
  • Publish at least 5,000 adjudicated mapping decisions, 2,000 hard negatives and 50 held-out source schemas or origins.
  • Deliver Visual Atlas Agent v1, proof-linked narration and at least five public semantic change streams.
  • Release publisher generation, validation, policy, correction and revocation tooling with a first external pilot.
  • Publish dedicated deployment, self-hosting and federation alphas plus an independent security-review reserve.
  • Reach a 90-day floor of 3 million real packets and 300,000 proof-linked frames; stretch to 5 million and 500,000.

Agent and TNOE admission path

The public agent and model program has its own gates because corpus size, semantic review, endpoint security, model spend and training readiness are not interchangeable.

R0 advanced

The protected candidate now includes a nine-tool MCP surface, one-call bounded agent context and Agent Answer 0.2 epistemic reconciliation. Public activation and paid frontier-model evidence remain R1 decisions.

  1. R0 — Agent-facing contract hardeningNine bounded MCP tools, including a compact task-ready agent context; evidence-bound Answer 0.2; effective query templates; proof escalation; readable telemetry; full Semantic Query schemas; and public explanatory surfaces. Admission: founder review required.
  2. R1 — Frontier agent pilotPublic read-only MCP, 20-scenario cost pilot, same frontier model across conditions, evidence-bound answers and published losses. Admission: budget, credential, public-endpoint and operator-security approval required.
  3. R2 — Ontology corpus and TNOE data admissionFive source families, 250 unique mapping units, 500 signed decisions, 1,000 reviewed hard negatives and four useful universes. Admission: human policy and signed semantic review required.
  4. R3 — TNOE-Align 0.1 training candidateDual-graph alignment model, concept and role ranking, relation classification, calibration, abstention, identity-risk and security-anomaly evaluation. Admission: R2 floors and source-family-held-out evaluation required.
  5. R4 — Continuous diverse Web expansionWorld State expansion, Crossref, APIs.guru, NVD and Federal Register; packet, frame, delta, diversity and query-utility releases. Admission: per-source policy and bounded evidence acquisition required.
  6. R5 — Powerful TNOE research lineReplace frozen text features when justified, add temporal change learning and active-review proposals, and evaluate generalization without granting model authority. Admission: measured data, compute, safety and independent evaluation gates required.

Agent interface Full model architecture AgentBench

Constitutional constraints

Funding cannot purchase semantic rank, source authority, canon admission, trust status, governance rights, or exclusive ownership of public work. Model output remains candidate-only. Production payments, write actions and arbitrary URL execution require separate future gates.