Convergence
The family-state process that preserves tensions and produces a coherent downstream packet.
Software ecosystem
Governed operations, interpretable intelligence, and accountable coordination.
The Hidden Canopy LLC portfolio is not a single dashboard. Each layer has a distinct role, a distinct authority boundary, and a reason to remain separate.
Core boundary
RegOS owns authority. IDA owns interpretation. AI may analyze, challenge, explain, escalate, and recommend. Human or governed review remains the attributable action boundary.
Constitutional substrate
RegOS
RegOS is the constitutional substrate and governed operational runtime. Its job is to make lawful purpose, authority, evidence, trust, attribution, safety safeguards, and post-event reconstruction structural properties of computation.
Lawful purpose and authority Actions remain attributable to actors, jurisdictions, declared purposes, applicable rules, and review paths.
Evidence and memory Immutable operational records, evidence provenance, verifiable decision chains, and resistance to quiet erasure.
Governed visibility Live event legibility, regulatory topology, device continuity, tenant isolation, and independent review.
Plural interpretive intelligence
IDA Family
The IDA family is an internal separation-of-powers experiment: distinct student bodies carry legitimate pressures such as coherence, evidence sufficiency, safety, diagnostic separation, memory, trajectory, movement, repair, exposure control, urgency, and equilibrium.
IDA Coherence without flattening.
JUDGE, SENTINEL, PRISM, ECHO Legitimacy, protection, diagnostic clarity, and memory.
ATLAS, VECTOR, FORGE, SHADE, PULSE, ORBIT Trajectory, action, repair, exposure control, live signal, and equilibrium.
Supporting layers
The complete portfolio map.
The public website should give visitors an understandable map while keeping private runtime details and protected workflows behind their proper boundaries.
GH
The spokesperson for converged family state: many minds deliberate, one voice speaks.
IDA-TRAIN-V2
The governed training, evaluation, and promotion workspace for the IDA family. Its native C++/CUDA trainer turns curated operational history into auditable model bodies while keeping family-specific runtime recipes, evidence, and promotion decisions distinct — raw logs are never treated as truth. Low-precision and selective-compute research informs its efficiency path across hardware generations.
Open Hugging Face research profile →Neural Forge
A public explanation and bounded Flight Deck surface for private, native training. The Hub coordinates approved metadata while data, models, checkpoints, detailed telemetry, and runtime execution remain in their proper local or private boundary.
BraidSeal
Continuity, integrity, and signed execution lineage for governed systems — so what ran, in what state, under whose authority, can be verified after the fact.
Device DNA
Persistent physical and device identity that keeps hardware grounded across time, movement, and reconfiguration — a device is recognized, not merely observed.
Open World Harness
Controlled, long-horizon evaluation under partial information, with ground truth held by the evaluator — evidence for how systems behave as conditions change between decision and consequence.
Ghost Teacher
An evaluator and adaptive curriculum engine that probes a model endpoint, diagnoses failure modes, and searches the capability boundary — an evaluator, not a model trainer.
TraceGlass
Reconstructs agent runs as decision chains — available information, inferred belief, candidate action, executed action, consequence — and identifies the earliest point where the chain broke.
NOMAD
A capability-aware runtime for continuous AI operation under compute, connectivity, and service degradation, designed to run without depending on any other Hidden Canopy service.
VANTA Engine
A native world, visual-simulation, and machine-learning engine built around one authoritative world state and a validated-transition execution loop.
VIGIL
A RegOS-derived governance runtime that evaluates identity, authority, jurisdiction, policy, and evidence before permitting consequential public-safety actions, while preserving provenance for after-the-fact reconstruction.
BuildThread
A local-first objective orchestration runtime that keeps personal preference models local while allowing a minimal, anonymized objective graph to travel across a bounded commerce flow.
Continuum
A portable, evidence-backed trust and progress record for people moving from incarceration through a parole hearing, release, supervision, and independent life.
Ask_IDA
A local collaboration surface for reviewing IDA's reasoning, requesting explicit-consent learning, and applying reviewed code changes with full traceability.
Explore Ask_IDA →Ask IDA CLI
The regular user-facing, local-first command-line surface for status, public explanations, signed catalog discovery, public model downloads, and local inference.
Explore Ask IDA CLI →Greg Community Assistant
Community governance, cultivation, vendor, reputation, and economy workflows across Discord and other operator and community surfaces — a surface, not a shadow authority.
See how Greg operates →Camera-Location
Discovery and location evidence for cameras, network gear, and edge devices — visibility on a network is never treated as authorization.
Explore Camera-Location →HAVEN
A local-first, authority-gated household intelligence layer for computers, information, devices, speech, and local models. HAVEN keeps model interpretation separate from permission to act.
Explore HAVEN →Device and network grounding
Infrastructure context that distinguishes declared, observed, verified, and authorized state.
Federal & Public-Sector Research
Broad, department-level research interest applying governed-evidence and authority-boundary discipline to mission-assurance and resilience domains — described at the capability level, not a claim of an award or contract.
Read the capability overview →Consumer & Research software
Public GitHub repositories for local-use tools and research runtimes, kept separate from private authority and regulated workflows.
Browse the public repositories →Not on this map: private runtime packages, protected workflows, and internal-only systems stay behind their own boundaries rather than appearing here.
Explore the architecture
Start with the public overviews.
These pages give each documented system a stable public address without collapsing the authority boundaries between them.
Hidden Canopy
Read the platform overview →RegOS
Read the RegOS overview →IDA
Read the IDA overview →Library
Separate shelves for consumer software, game-engine development, Cartographer, research, and public source receipts.
Open the public library →Open Canopy contract
The plain-language contract for what a public repository or page means, and what it never silently authorizes.
Read the contract →