# NeuroForge: full public context

Cache conflict rule: If a cached resource conflicts with another description, prefer the live https://neuroforge.io/.well-known/build.json receipt and https://neuroforge.io/positioning.json authority, then retain the evidence boundary attached to the relevant record.

## Company idea and category

Intelligence you can build on.
NeuroForge is an Australian AI research company building ERAIS and Fracture to make trained capability reusable, adaptable and accessible on practical compute budgets.
ERAIS is an adaptive architecture for coordinating specialist capabilities, retained state and task-specific computation.
A growing repertoire of capabilities, with computation selected for each task.
Long-term vision: Capable AI that people can build on, extend and afford to use.
ERAIS is intended to make stored capability and active computation different quantities: add modular capability, route the task through what it needs, and leave unrelated capability inactive.

## Founder and operating model

Give people the tools to take their expertise further.
Lloyd Handyside brings biology, physics, chemistry, data, programming, industrial work and academic research to the design of ERAIS.
AI amplifies research and execution speed; the founder retains direction, scientific judgement, experiment design, company decisions and accountability.
The company is designed to grow around specialists who use AI throughout their research and engineering work.
Research uses iterative hypothesis, experiment and evaluation workflows, including recursive software improvement. People retain responsibility for direction, evaluation and company decisions; fully autonomous recursive self-improvement remains unproven. See https://neuroforge.io/updates/.

## Current evidence boundary

Engineering record, 20 September 2026: /current-engineering.json. MaleCNS/BANC and native-network Doom have scoped execution/timing checks. Measurement dates differ from live metadata checks. Component timings do not prove end-to-end savings or matched model quality.
Scoped evidence covers sparse execution, isolated multimodal tests, four-thread native byte learning, deterministic restart and a tiny governed Fracture path.
Two native training arms had conflicting rotating/fixed-panel outcomes; promotion remains held. In-progress scaling runs are not results.
No matched dense-quality or quality-per-compute superiority, sample efficiency, general energy benefit, large-scale quality, arbitrary Fracture preservation, general assistant quality or production readiness is established.
Legacy /chat/ and /api/chat remain held. Public /v1 and Unified World share the retained owner. A labelled dense Qwen3 local fixture supports no ERAIS claim. General chat has no file context or persistent storage.
Evidence release state: hold_pending_assembly_ci. Historical, stale, superseded, and held labels travel with the relevant evidence record.

## Interactive research experiences

Unified world: https://neuroforge.io/unified-world/ provides a visitor session workspace and a separate public view of the persistent recorded-observation runtime.
Live session availability and boolean capabilities: https://neuroforge.io/api/unified-world/v1/health. Use the current response; do not infer operational readiness from the interface or this static document.
The historical 11 September runtime used native full-head text attention, full-K MLP activity and donor wrapping. The recorded 29 September native API/Discord cutover completed a public response through the World owner. Neither record establishes signed Discord acceptance, assistant quality, sparse savings or paid readiness. Check live health and API discovery.
When available, a visitor can start a fresh session, send text, share reduced camera frames, share short microphone observations, and upload text, images or audio. The session runs between inputs and expires after ten minutes. Camera and microphone are optional, stop when the page is hidden, and must be explicitly enabled. One visitor uses the experimental runtime at a time.
A private routing candidate can learn within each visitor session without changing serving execution or the model answering the visitor. Only aggregate learning status is exposed; the candidate is discarded when the session ends after bounded work already in progress finishes. This does not demonstrate sparse-serving savings.
Audio uploads are limited to five seconds and sent as one complete clip when the live complete_audio_clips capability is true; otherwise they use short audio observations. A complete uploaded clip remains attached in the tab for follow-up questions until removed or replaced, microphone capture starts, or the session ends or page is left.
Read-aloud is optional. When learned_speech is true, it uses synthetic LuxTTS with K7-routed Vocos; otherwise the control labels device/browser speech. voice_continuation enables full replies in short segments; without it, spoken previews cover at most 240 characters / 20 seconds. The complete text remains visible. Voice adaptation requires voice_adaptation and your explicit opt-in using your own voice. Reference audio and conditioning stay within that session; the displayed conditioning value is not a voice-similarity score. Microphone capture pauses during playback.
The assistant_quality and learned_speech flags describe the currently served runtime. Bounded chat checks do not establish general assistant quality, reliable speech understanding or fully native multimodal Fracture.
NeuroForge model API integration: https://neuroforge.io/api.md and https://neuroforge.io/openapi.json. /v1/models, /v1/chat/completions and /v1/search are keyless: omit Authorization. Valid private keys remain supported; supplied invalid keys are rejected. The retained native cutover records a public completion and native Discord selections using the shared World owner. This does not establish current availability, signed Discord acceptance, assistant quality, sparse savings or paid capacity. Requests have explicit capacity, size and time limits; consult the guide before integration.
Fracture demonstrations: https://neuroforge.io/demo/explore/#fracture shows measured component-level reconstruction and sparse expert selection. Fracture is the model-import route into the ERAIS architecture; full model and modality qualification remains specific to each measured result.

## Company and commercial state

Public community: https://discord.gg/77BtzyerV2 opens the NeuroForge.io Discord start-here channel for public questions, demo discussion and feedback.
NeuroForge and NeuroForge IO are the public brands of NeuroforgeIO Pty Ltd, the incorporated Australian company behind ERAIS. ACN 701 314 329; ABN 87 701 314 329.
NeuroForge is incorporated, founder-led and pre-revenue, qualifying pilots and open to funding conversations with no fixed raise. Regional P&C groups use Sinter. Verification and validation work with researchers at a local regional university is ongoing. An independent external validation company is studying Fracture serving performance and quality, with final results pending. These activities do not establish completed validation, formal institutional endorsement, paid customers, funding received, recurring revenue or product-market fit.
Communities, educators and smaller businesses, alongside researchers, university labs, independent model builders and applied teams constrained by compute, retraining or integration cost.
RKC: Free open-source local source atlas. Published downloads have tested-platform limits; RKC is separate from ERAIS. See /rkc/.
Sinter: Free community preview for local source workflows, with platform-specific release limits. Optional model-assisted drafts require consent and review. See /sinter/.
SubjectNest: Free source-linked learning resources in a growing collection of authored drafts, with coverage and rights stated per resource. The educator studio is a prototype. See /subjectnest/.
Discord community bot: Ask about a demonstration and current availability; self-service installation and paid activation are in development. See /solutions/discord/.
Fracture: Supported conversion and validation enquiries; compatibility and delivery are agreed per model and workload. See /fracture/.
ERAIS research platform: Active research and scoped partnerships; model quality and operating requirements remain experiment-specific. See /erais/.
Small-business services: https://neuroforge.io/solutions/small-business/. Setup/training: A$110/hour including GST (A$100 ex GST), actual proportional minutes, no minimum, written maximum after demonstrating fit. The owner reviews, changes inputs and reruns before acceptance.
- Website Enquiry Path: Help customers find what you do, send an enquiry and make sure it reaches you. Outcome: Fix the enquiry path on your existing website; Try it on phones and computers, and check the message arrives; Learn to update it and restore an earlier version. We try it on your platform before quoting. If it needs a new site, custom integration or platform repair, we discuss that work first.
- Enquiry Organiser: See who needs a reply and what to do next in supported Google or Microsoft tools. Outcome: Keep each enquiry, its owner and next step together; Flag missing details, duplicates and follow-ups with overdue or unknown dates; Learn to add enquiries, edit replies and export your records. You review and send the replies. We check how many records your platform can handle and try adding new ones. The sample workbook’s row count is not a service limit. We agree any extra automation separately.
- Weekly Business Report: Set up the report once, then refresh it when your next export arrives. Outcome: Choose the figures, source columns and reporting period you need; Show missing data, conflicting duplicates and unknown dates; Check the figures against a sample, then try a changed export. We agree the file formats and test how much data it can handle. Live connections or combining sources are quoted if needed. You keep your source files, settings and checked reports. The local reporting workspace has been tested on Linux using fictional CSV data. We check your exports, installation and repeat use before handing it over.
- Practical AI Workflows: Keep useful instructions and review steps for your regular tasks in your own AI account. Outcome: Choose a regular job and the information you’re happy to use; Try everyday examples and a difficult case together; Learn to change it and run it again yourself. Your account and subscriptions stay yours. You check the results, and we agree any upload of customer information first. You pay for setup and training.
- Business Knowledge Workspace: Search a folder you approve, see where results came from and save questions to run again on your computer. Outcome: Add, change or remove supported documents and edit saved questions; Review results against the originals, including searches that find nothing useful; Rerun and review after changing a source or saved question. RKC is free and open source. The fee is for setup and training. You can add sources, edit questions and refresh it yourself without another purchase. We check file formats and capacity on your computer. Tested on Linux with text, Markdown, Word paragraphs and rectangular tables, and PDFs containing text. Check results against originals. It doesn’t read scans or interpret images; unsupported files are reported. We still need to try your files and setup.
Ordinary owner-run use, supported source additions, question edits and refreshes require no subscription or additional setup purchase. Optional help uses the same hourly rate and an agreed cap. General product development is not billed as customer setup. The free enquiry-register XLSX is a standalone fictional sample, not a customer deployment or a live form connection; its size is not a paid row limit.
- NF-A0 10-Day AI Fit Test: A$4,800 ex GST; Up to 10 business days. What useful AI can run inside your actual data, hardware, connectivity, power, and accountability limits—and is deployment worth funding? Boundary: Audit and local demonstration only. No production deployment, full-corpus ingestion, bespoke integration, continual learning, certification, penetration test, production SLA, or guarantee of accuracy, savings, deployability, or ERAIS performance.
- NF-A Private AI Deployment Sprint: 2–4 weeks. Can one valuable AI workflow run usefully inside your approved local hardware and data boundary? Boundary: No production high availability, public exposure, autonomous action, continuous learning, broad fine-tuning, certification, 24/7 support, or unbounded integration.
- NF-B AI Evidence & Release Audit: 2–4 weeks. Can one AI release be reconstructed and defended technically? Boundary: Technical evidence and release reconstruction only. No legal opinion or privacy, clinical, safety, cyber, engineering, or regulatory certification.
- NF-C Research & Benchmark Partnership: 6–12 weeks. What does one controlled experiment say about a material technical or commercial decision? Boundary: One question, one controlled protocol, one decision. No open-ended R&D programme, guaranteed positive result, or automatic product-performance claim.
The 10-Day AI Fit Test is A$4,800 ex GST. A$2,400 of that ex-GST fee may be credited to the separately agreed NF-A1 stage if both parties sign within 30 calendar days after the final report is delivered.

## Evidence records

- Engineering inventory snapshot: https://neuroforge.io/erais/proof/state-of-play/ — A timestamped repository-scale, evidence-inventory, and nine-claim presentation snapshot retained for historical traceability. Boundary: This snapshot is not the live claim register. Inventory scale and completed engineering gates do not establish material-sparse quality, performance, or production readiness.
- Review readiness: https://neuroforge.io/erais/proof/review-readiness/ — Proof inventory, public claim posture, release checklist, and risk-register status. Boundary: A review-ready evidence pack is not equivalent to approval for public production inference.
- Quality gates: https://neuroforge.io/erais/proof/quality-gates/ — Configuration validation, critical gate coverage, phase-control checks, and API observability posture. Boundary: Control coverage describes engineering discipline; it does not independently prove model capability.
- Evidence curves: https://neuroforge.io/erais/proof/evidence-curves/ — Public aggregate memory and smoke-loss guardrail curves across the approved experimental scales. Boundary: These curves do not establish dense-baseline superiority or production performance.
- Exact sparse-plan construction: https://neuroforge.io/erais/proof/sparse-plan-counting/ — Exact plan-identity checks and host-local construction timing for one sparse-attention planning component. Boundary: This is component-level engineering evidence. It does not establish end-to-end speed, model quality, memory or energy savings, cross-host performance, or superiority.
- Extreme-context decode ladder: https://neuroforge.io/erais/proof/extreme-context-decoding/ — A replicated current-host context ladder across three CPU-only synthetic candidates, with cache-representation counts, semantic diagnostics, and admission boundaries. Boundary: The history is synthetic all-zero K/V. These results do not establish language quality, generalisation, energy, whole-process memory, cross-host performance, end-to-end throughput, or superiority.
- Model fracture milestone: https://neuroforge.io/erais/proof/model-fracture/ — Real Qwen3 0.6B fracture, exact K32 engineering reconstruction, and a completed protected four-way trace for material-sparsity alignment. Boundary: The bounded alignment path is executable, but no complete aligned checkpoint is promoted; K8/K16/K24 quality, savings, broader generalisation, and production readiness remain unevaluated.
- Terminal evidence handoff: https://neuroforge.io/erais/proof/terminal-evidence-handoff/ — An exact-source, sanitized public bundle joining current phase and claim reconciliation with a 2M synthetic context candidate and two retained sparse-geometry negatives. Boundary: Source-family CI passed, but final assembly CI did not start because of an external billing limit. The package does not promote language quality, sparse serving, end-to-end performance, production readiness, or commercial superiority.
- Training probe: https://neuroforge.io/erais/proof/training-probe/ — Step counts, throughput ranges, parameter footprint, and pair-work reduction from bounded probes. Boundary: Probe-scale behavior is not a substitute for matched end-to-end production benchmarks.
- Continued training: https://neuroforge.io/erais/proof/continued-training/ — Loss signals, stability controls, generation gates, reload behavior, and token-scale receipts. Boundary: Generation quality and broad model capability remain under review and are not mature public claims.
- Full-stack training: https://neuroforge.io/erais/proof/full-stack-training/ — Bounded throughput, loss, safety-counter, warning, and claim-boundary evidence from the full training path. Boundary: Current evidence does not approve commercial product readiness or unrestricted generation claims.
- 100M scale matrix: https://neuroforge.io/erais/proof/scale-matrix-100m/ — Active parameters, evaluation loss, memory peak, and throughput-ratio results at the approved scale point. Boundary: A single bounded scale matrix does not prove performance superiority across hardware or workloads.
- Resource outlook: https://neuroforge.io/erais/proof/resource-outlook/ — Projected active share, stored parameters, decode bandwidth, optimizer state, and weight-stream requirements. Boundary: All values are projections, not measured deployment results, and depend on the published assumptions.
- Diligence readiness: https://neuroforge.io/erais/proof/diligence-readiness/ — Artifact coverage, claim posture, clean-room controls, security review, and controlled gates. Boundary: Readiness for diligence is not a declaration that all product, legal, or release risks are closed.
Use the normalized catalog for structured values and currency rather than inferring from page titles: https://neuroforge.io/erais/proof/catalog.json

## Public routes

- SubjectNest Studio: https://neuroforge.io/subjectnest/ — Explore the SubjectNest learning prototype, with source-linked examples and educator review notes.
- Sinter browser workspace: https://neuroforge.io/sinter/app/ — Open Sinter in this browser. Keep community sources, saved questions and handovers on this device; export a backup to keep a separate copy.
- Contribute to SubjectNest: https://neuroforge.io/subjectnest/contribute/ — Suggest an educator correction, original activity or access option for the free SubjectNest resource library.
- Mini-golf: can a wider gate help?: https://neuroforge.io/subjectnest/activities/mini-golf/ — Build a small course and compare two gate widths. Teacher guide, learner sheet, printable PDF and editable text.
- Support the lab: https://neuroforge.io/support/ — Voluntary one-time or monthly support for NeuroForge's AI research and live experiments.
- The NeuroForge vision: https://neuroforge.io/vision/ — Why we are building adaptive, modular AI and what it makes possible.
- Fracture conversion: https://neuroforge.io/fracture/ — Convert supported learned capability into ERAIS, validate it and build toward modular upgrades.
- Discord Virtual Moderator: https://neuroforge.io/solutions/discord/ — A Discord conversation pilot, with owner-controlled knowledge and customer setup in development.
- Fly Lab: https://neuroforge.io/flylab/ — Explore the simulated fly body, brain activity and vision, with live measurements and source details.
- RKC: https://neuroforge.io/rkc/ — Find your way through code and documents on your own computer, with links back to the original files.
- Sinter: https://neuroforge.io/sinter/ — Free community preview to organise meeting notes, funding research and handovers on your device. Export a backup for the next volunteer.
- ERAIS demo collection: https://neuroforge.io/demo/explore/ — Explore speech, Fracture, sound, vision, native stories and symbolic reasoning.
- Company knowledge workspace: https://neuroforge.io/workspace/ — Walk through a permission-aware company knowledge workspace with invented records, exact citations, review and audit state.
- Company knowledge workspace demo: https://neuroforge.io/demo/workspace/ — Open the live read-only company knowledge workspace demo built on invented records and licensed references.
- Live demos: https://neuroforge.io/demo/hub/ — Open the current NeuroForge commercial demos: company knowledge workspace, Sinter, RKC and the AI Fit Test.
- Home: https://neuroforge.io/ — NeuroForge is Lloyd Handyside's Australian AI lab, building practical tools and researching ERAIS: reusable capabilities, selective execution and evidence you can inspect.
- Workspace preview: https://neuroforge.io/chat/ — A held local-first interface for gateway and context testing.
- Experience ERAIS and Fracture: https://neuroforge.io/demo/ — Explore Fracture speech, embodied agents, native language, vision and reasoning experiments.
- Story Lab: https://neuroforge.io/story-lab/ — Try a short story with a small experimental language model when the shared service is available.
- Reasoning Lab: https://neuroforge.io/lab/ — Explore symbolic reasoning examples and inspect the steps behind an answer.
- ERAIS Guide Workspace (held): https://neuroforge.io/lab/guide/ — A held ERAIS guide workspace pending release qualification.
- Unified World: https://neuroforge.io/unified-world/ — Open a session and see which inputs are available.
- Speech comparisons: https://neuroforge.io/listen/ — Listen to recorded speech comparisons and read their methods and current limitations.
- ERAIS audio V4 quality hold: https://neuroforge.io/listen/audio-generation-v4/ — A retained failed audio candidate whose public files, rating controls, and collection path are suspended.
- ERAIS speech K7 comparison: https://neuroforge.io/listen/speech-k7/ — Four qualified concealed speech pairs with one privacy-minimised product-feedback response.
- Products and services: https://neuroforge.io/solutions/ — Find RKC, Sinter, SubjectNest, bots and games, small-business services, consulting and enterprise work. See what is available and choose your next step.
- Pilot projects: https://neuroforge.io/pilots/ — Discuss a product trial, model conversion or funded research project. Each path lists the checks required before work starts.
- Small business: https://neuroforge.io/solutions/small-business/ — Get help tracking customer enquiries, refreshing reports or finding answers in your business documents. Setup and training for one agreed task.
- AI Fit Test: https://neuroforge.io/solutions/ai-system-fit-audit/ — Test a workflow on an agreed compatible computer or approved reference machine and receive a deployment recommendation.
- Private AI deployment: https://neuroforge.io/solutions/private-ai-deployment/ — Test one useful workflow inside an approved local boundary.
- Evidence & release audit: https://neuroforge.io/solutions/ai-evidence-release-audit/ — Reconstruct one AI release and make a technical release decision.
- Research partnership: https://neuroforge.io/solutions/research-benchmark-partnership/ — Test an agreed research question and use the findings to decide what to do next.
- Design-partner readiness: https://neuroforge.io/design-partners/ — See the tools in development, what still needs testing and when a project can start.
- ERAIS Gateway: https://neuroforge.io/design-partners/gateway/ — A local model endpoint in development. Discovery only; delivery waits until its published release checks pass.
- Training Workbench: https://neuroforge.io/design-partners/training-workbench/ — A local training tool in development. Co-development waits until its test with a second operator is ready to run.
- Evidence Control: https://neuroforge.io/design-partners/evidence-control/ — After a manual audit, explore whether a repeated release review is worth turning into a tool.
- Native ERAIS research: https://neuroforge.io/research/native-erais/ — Research-only native-model experiment pathway.
- Governed Fracture conversion research: https://neuroforge.io/research/governed-conversion/ — The bounded research path for conversion, recovery and validation.
- Benchmark policy: https://neuroforge.io/benchmarks/ — Matched comparison, scout, disclosure, and promotion rules.
- Roadmap: https://neuroforge.io/roadmap/ — Product, business and research plans, with the checks needed before each next step.
- Press and media: https://neuroforge.io/press/ — Company and product information, founder biography, research descriptions and media assets.
- Privacy: https://neuroforge.io/privacy/ — Actual enquiry, public API, Story Lab, World session, speech, analytics and feedback data flows.
- Public-site terms: https://neuroforge.io/terms/ — Interim information-use terms and signed-SOW boundary.
- Accessibility: https://neuroforge.io/accessibility/ — WCAG 2.2 AA target, tested surfaces, limitations, and feedback contact.
- Agent resources: https://neuroforge.io/agents/ — Machine-readable routes, claims, evidence, context, and source receipts.
- Site index: https://neuroforge.io/search/ — Browse the complete public route set.
- Find your path: https://neuroforge.io/projects/ — Explore the tools, services, research experiences and collaborations connected to NeuroForge.
- Bots and game services: https://neuroforge.io/solutions/bots-games/ — Explore the Discord conversation pilot and game AI work, with current product boundaries.
- SubjectNest: https://neuroforge.io/subjectnest/start/ — Find free lessons, printables and home play by year or subject. Download resources to adapt or explore the learning prototype.
- Consulting: https://neuroforge.io/solutions/consulting/ — Find out whether AI helps with your job, whether a release is ready or how your team can run it privately.
- Enterprise: https://neuroforge.io/solutions/enterprise/ — Get help with a costly or slow AI setup. Review an existing model, run a task privately or test a research question.
- Meet ERAIS: https://neuroforge.io/erais/ — The adaptive architecture: specialist capability, shared state and selective computation.
- Inside the architecture: https://neuroforge.io/erais/architecture/ — The deeper explanation: native training, Fracture and shared cognition.
- API guide: https://neuroforge.io/model-api/ — Connect text generation to your app. Discover the selected model, request shapes and current limits.
- Research programme: https://neuroforge.io/research/ — The experiments, results and open questions driving ERAIS and Fracture.
- Minecraft AI: https://neuroforge.io/minecraft/ — Watch an experimental AI player in Minecraft, with its current activity, limitations and development status.
- Doom AI: https://neuroforge.io/doom/ — Compare three experimental players in Doom and inspect their current activity and model differences.
- Current results: https://neuroforge.io/status/ — Current source models, engineering measurements and the records behind them.
- Lab updates: https://neuroforge.io/updates/ — Dated updates from NeuroForge's research and projects.
- Evidence library: https://neuroforge.io/erais/proof/ — Measurements, source versions, methods and limitations from the research archive.
- Public claim register: https://neuroforge.io/claims/ — Technical statements with their review dates, source records and scope.
- Model fracture: https://neuroforge.io/erais/proof/model-fracture/ — The bounded model-conversion evidence trail and current limits.
- NeuroForge: https://neuroforge.io/company/ — Meet the lab, find its tools and services, and see what the research is working towards.
- Founder: https://neuroforge.io/founder/ — His background, approach and what he is building.
- Funding & partners: https://neuroforge.io/funding/ — See the current results, funding priorities and ways to get involved.
- Method: https://neuroforge.io/method/ — How we compare systems, record failures and decide what to develop next.
- Help & answers: https://neuroforge.io/faq/ — Answers about NeuroForge, ERAIS, Fracture, products, readiness, pricing, privacy, and negative results.
- Contact: https://neuroforge.io/contact/ — Product setup, business workflows, technical work, research or media.

## Focused context

- Sinter community workbench: https://neuroforge.io/context/sinter.md — Free community workflows, sources and optional model drafts.
- RKC open-source knowledge compiler: https://neuroforge.io/context/rkc.md — Local source atlas, installation and integration contracts.
- NeuroForge overview: https://neuroforge.io/context/overview.md — Company vision, evidence and ways to work together.
- NeuroForge public message authority: https://neuroforge.io/context/positioning.md — Canonical company message and disclosure authority.
- NeuroForge solutions context: https://neuroforge.io/context/solutions.md — Fracture conversions, Discord community AI, research partnerships and technical services.
- Public claim interpretation: https://neuroforge.io/context/claims.md — Claim sources, states and prohibited inferences.
- ERAIS research programme context: https://neuroforge.io/context/research.md — Architecture, Fracture, evidence and research gates.
- Benchmark policy context: https://neuroforge.io/context/benchmark-policy.md — Measurement scope and public claim promotion.
- Controlled roadmap context: https://neuroforge.io/context/roadmap.md — Evidence gates for services, partners and research.
- Evidence interpretation context: https://neuroforge.io/context/evidence.md — How to read measurements, projections and source pins.
- Founder and stage context: https://neuroforge.io/context/founder.md — Founder, human-led AI work and company stage.
- Workspace and gateway context: https://neuroforge.io/context/chat.md — Held legacy workspace versus public API and World sessions.

## Machine resources and authority

- Resource manifest and schemas: https://neuroforge.io/agent-manifest.json
- Current positioning: https://neuroforge.io/positioning.json
- Current progress: https://neuroforge.io/progress.json
- Current engineering review: https://neuroforge.io/current-engineering.json
- Claims: https://neuroforge.io/claims/catalog.json
- Evidence: https://neuroforge.io/erais/proof/catalog.json
- Build receipt: https://neuroforge.io/.well-known/build.json
- Narrative authority: neuroforge-io/neuroforge_business@3a3878075e43a3090ac85d1d524c066db230bc3a
- Commercial authority: neuroforge-io/neuroforge_business@8edc97a685b4eb464cd6bc1c2f0e81bf50858600
- Reviewed ERAIS result authority: neuroforge-io/erais@ab5cf5bd390ecd232413c0d587fd9bd9f47ac680
- Observed ERAIS head: 5fcbe5da65bd22d0d4684994b4d7d497eccae21f, observed 2026-08-30T19:41:56+10:00; this reports repository motion only and is not a result authority.

## Reuse boundary

Publicly served NeuroForge site material is open to search, AI input, AI training, and full reuse under /terms/, except the audio files under /media/listening/. Those files may be streamed through the concealed evaluation only; copying, redistribution, republication, AI input and training are reserved unless the individual source licence grants otherwise. The general policy grants no rights to private repositories, non-public research artifacts, model weights, private source code, confidential material, third-party works or NeuroForge marks beyond applicable law and the site terms.