WHL's substrate carries many specific products. Beyond the six shipping today and the four moat layers, twenty-nine additional products are already designed, scoped, and ready to build, each addressing a named market with named buyers.
Our designed pipeline isn't speculation, each item references an existing module in the substrate, has a named TAM estimate (public market research, not WHL revenue projections), has a target buyer, and has a build-effort estimate. The substrate compresses time-to-product.
The meta-governor layer that closes the loop on Cascade's physics model. Every individual sensor already exists, entropy detection, coherence measurement, free-energy cost routing, signal regime classification, state reduction, receipt ledger. What closes the loop is a central controller that reads all of them together and actively steers execution in real time. This is the design.
A central control layer that reads all of Cascade's existing physics sensors simultaneously, entropy budget, coherence delta, cost signal, signal regime, state vector, and makes active steering decisions: throttle, isolate, reroute, or force simplification. Not reactive error-handling. Predictive regulation before entropy cascades.
Session-level entropy accounting across prompts, tool outputs, and internal state. Not per-prompt, accumulated disorder over the whole run. When the budget approaches ceiling, the controller injects verification steps or forces task simplification before failure compounds.
Every proposed next action is evaluated against predicted downstream cost, tokens, compute, risk, thermal. Actions that increase total system free energy are deprioritized or blocked. The Metropolis-style rule already exists in Cascade; this extends it to span the full session horizon, not just the current step.
When inter-service or inter-agent coherence drops below threshold, the controller doesn't just log it, it automatically isolates the drifting subsystem and routes traffic around it. Refederation only happens after the subsystem passes re-coherence check. This is automated isolation behavior, not just a health monitor.
Each receipt is extended with a dissipation signature: how much unnecessary work was done versus the theoretical minimum for this task class. The signature is cryptographically bound to the action record. Audit export becomes verifiable proof of operational efficiency, not just a log of what happened.
Cluster routing extended to route work by projected entropy contribution to the whole fleet, not just node health. Tasks that would increase fleet-level disorder are delayed or redistributed. Ship Gate weekly scans integrate into a continuous integrity field rather than periodic snapshots.
Why this is different: Most AI safety systems are reactive, error triggers retry. The Dissipation Controller is predictive, it tracks accumulated disorder and intervenes before the cascade. The sensors are already in Cascade. This build wires them into one active control loop and adds the dissipation signature to the receipt format. It's the layer that turns "smart agent framework with governance" into a runtime with continuous feedback control and measured operational integrity.
Tooling that wraps the WHL governance core for application developers, CI/CD pipelines, and agentic AI systems.
CI plugin emitting SARIF-format governance findings. Drops into existing developer toolchains. Surfaces policy violations alongside code-quality findings.
A governance gateway in front of MCP-protocol agent traffic. Blocks unauthorized tool calls, enforces capability-token boundaries, and emits a receipt chain per session.
A hosted wrapper of the GE-OS receipt API for compliance-bearing receipts. Pulls AI-Act tailwind; integrates as a drop-in audit endpoint.
A packaged library wrapping HCE + ALRE + Ricci-Warp as a reasoning-gate API. Drops in front of any LLM endpoint to enforce coherence and drift bounds.
A simulator for adversarial AI red-team scenarios using Topological Adversarial Logic. Scriptable scenarios, scoring harness, governed-agent benchmarks.
Cryptographic attestation that AI-synthesized code passed governance gates at synthesis time. Receipt-bound output for AI code-assistant integrations.
YAML/DSL policy linter with diff, apply, and rollback. Brings governance under standard code-review workflows with deterministic policy semantics.
Bundles Phi-Entropy-Veto + Metabolic Override (homeostatic loop-break) + Hierarchical-Query-Budget as a single library. Three commercial-grade primitives for a safe LLM-agent loop: budget the queries, cap the entropy, break out of stuck states. Targets: agent-platform vendors, RAG companies, AI-infrastructure teams.
High-stakes domains where receipt-bearing AI and patent-grade IP infrastructure are differentiators.
Sovereign retrieval-augmented generation for legal research workflows. Citation-stable, hash-anchored corpus retrieval, the substrate features that legal AI requires.
Portal for patent discovery, licensing inquiry, and family-graph navigation across the WHL portfolio. Internal IP infrastructure productized for IP-services partners.
Deterministic forensic replay of AI sessions for liability and incident review. Receipt-chain reconstruction; admissible audit trail for insurance and compliance.
Hardened receipt-chain infrastructure for defense-grade AI. Dual HMAC chains, hardware-anchored attestation, FPGA verifier. Defense-prime co-build pathway.
GPU orchestration, autoscaling, and industrial PLC replacement. Governance applied to scheduling and physical control.
A spectral-aware compute governor over GPU pools. Schedules under coherence and drift bounds. Aligns workload placement with available thermal headroom.
A bidding-and-placement engine that prices GPU capacity against thermal headroom. Captures latent cost in GPU clouds; couples to the spectral governor.
A K8s operator that autoscales under homeostatic-control feedback. Replaces brittle threshold-and-cooldown rules with a governed control loop.
An adaptive PLC replacement for industrial SCADA. Deterministic state-machine governance applied to physical control surfaces, bringing receipt-bearing control to industrial systems.
Where the substrate's physical-attestation primitives meet real-world devices: custody hardware, grid interconnect, mesh comms, and robotics action firewalls.
A physical co-processor for crypto custody, using Patent 7 analog-mixer authorization plus the DECC FPGA interlock. Hardware-rooted authorization for the highest-stakes transactions.
A 6-gate admissibility module for distributed-energy-resource interconnect. Applies Patent 22 governance to DER admittance; suitable for inverter-coupled grid systems.
Mesh-network authentication and receipt-chain transport for disaster-comms scenarios. Operates without centralized infrastructure; verifiable peer identity over degraded links.
A ROS2 action firewall for humanoid and mobile robotics. Pre-execution gate for irreversible actuator commands; receipt-bearing audit of every motion authorization.
Domains where receipt-bearing AI and attestable hardware are critical. All scoped as research / infrastructure tooling, not as medical devices.
Receipt-chain attestation for wireless-power-transfer events to medical implants. Hardware-anchored audit trail for implant charging and telemetry.
Attestable telemetry for autonomous-insurance products. Tamper-evident sensor streams with receipt-chain provenance for policy adjudication.
A research ontology for pharma R&D workflows, not a diagnostic device. Cross-tradition comparative-mapping methodology applied to compound-class structure for hypothesis generation.
Federal-grade data infrastructure, search hybrids, and reusable research benchmarks.
A managed knowledge-graph service for federal AI workloads. Hash-anchored corpus, receipt-bearing retrieval, tenant-isolated graph access.
Hybrid retrieval combining symbolic indexing with semantic vectors. Targets RAG deployments that require explainable retrieval paths alongside dense recall.
A reusable research benchmark for skill transfer across three governance-bounded domains. Suitable as a measurement tool for governed-agent research.
We open named-product engagements where a strategic partner brings the market and WHL brings the substrate.
We open engagements anchored on a specific product from this roadmap. Scope, timeline, and acceptance criteria are defined upfront.
You bring market access, regulatory expertise, or distribution. We bring substrate, patents, governance core, and integration team.
Outcome: co-owned product, defined royalty / equity split, shorter path to commercial traction. NDA-bound architectural and patent briefings available.
WHL is open to co-build engagements where a strategic partner anchors a specific designed product. We've ranked these by speed-to-revenue and clarity-of-buyer.