Public whitepaper · v2
The framework.
The boundaries.
A public account of LUMA Quant’s analytical direction, established architecture, research ambitions and evidence standards. Enough to understand the work without disclosing its proprietary implementation.
The August 2026 v2 edition is retained as the public framework overview. Product observations below refer to the addendum. The editorial update reorganises the content and adds a high-level research overview; it is not a new product audit or research-results release.
Dated product addendum
Product context · 19 October 2026
LUMA Quant brings research, cooperative leisure and token-related participation into one developing ecosystem. Each has a distinct entry point and purpose:
- AI Platform : the research and analysis interface for the current framework and experimental benchmark.
- Playground · LUMA Rooms : cooperative quizzes, word groups, clues, estimate rounds and optional duels.
- Token Portal : LUMA tokens, LUMAKeys and the NFT Forge, with separate user-approved purchase and mint actions.
The 19 October product-status record combines read-only public observations with dated first-party release evidence from 17–19 October. It did not include a new purchase, mint or credit-redemption test. Section 11 preserves those product distinctions. Current availability and transaction terms must be checked in the relevant application.
1. Executive summary
LUMA Quant is an independently developed analytical framework for pattern recognition and candidate regime detection under uncertainty. It studies whether structures persist, decay, conflict or reorganise across different analytical perspectives. No single score or model consensus is treated as proof.
Its first public reference implementation uses EuroMillions as a severe stochastic proof of concept. This provides a context for evaluating analytical behaviour with limited explanatory information, not a claim that randomness has been overcome.
2. Built under uncertainty. Designed across domains.
The ambition is to develop reusable principles for observation, memory, evaluation and controlled knowledge integration. A new domain requires its own data contract, baselines, implementation and validation. Transferability is a question to test, not an entitlement granted by an architectural diagram.
3. The extreme-uncertainty proof of concept
The EuroMillions benchmark has an observable outcome history and little privileged information about future outcomes. It makes disagreement, unstable patterns and post-outcome failures available for examination.
- Observable evaluation: compare earlier analytical states with later outcomes.
- Explicit uncertainty: retain conflicts, null findings and failure evidence.
- Limited information: do not infer a causal signal from historical structure alone.
- Separate evidence classes: distinguish synthetic demonstrations, retrospective analyses and prospective tests.
The proof of concept does not establish an independently demonstrated prospective EuroMillions prediction advantage. It does not make random events deterministic.
4. Recorded operational foundation
The operator’s roadmap records P0–P2 as implemented foundations: the first SaaS architecture, multi-axis analysis, regime-aware workflows, user access, analysis credits, report delivery, data pipelines and ENMF memory. These development statements describe the reported system scope. They are not an independent certification of reliability, completeness or predictive performance.
The technology roadmap distinguishes that foundation from ongoing product and research work.
5. Pattern recognition and regime detection
The framework examines candidate changes in stability, drift, interaction, agreement and structural persistence. Disagreement and decaying relationships are retained for evaluation. A candidate regime is a model hypothesis about behaviour, not an established property merely because it has a label.
6. Multi-axis analysis
Analytical perspectives separate stability, change, interaction, proximity, negative evidence and ranking before considering how they relate. Apparent confirmation can result from shared data or common assumptions, so multiple perspectives do not automatically provide independent evidence.
The research objective is to test whether a proposed structure survives appropriate controls and later evaluation, and to record the conditions under which it does not.
7. ENMF and experience memory
The Emergent Neuronal Memory Framework organises observations and analytical context into interacting memory roles. The existing public description includes source and temporal memory, change and interaction memory, negative-space evidence, entity context, ranking, decisions, post-outcome evaluation and a consolidated Belief State.
The Experience Layer is intended to connect research knowledge with actual model-run experience. A useful record includes provenance, context, uncertainty, version, applicability and counter-evidence. Keeping information does not establish that the system learns from it effectively.
Connections between ENMF, Experience Memory and the Belief State are evaluated separately from the presence of individual memory components. Production changes require additional engineering checks and explicit approval.
Explore the public memory overview
8. Platform architecture and orchestration
The existing SaaS architecture separates interfaces, access, services, analysis orchestration, data processing and storage. The public enterprise diagram is retained as the established architecture view; it is not a live inventory of the expanded research network.
The LUMA Orchestrator provides the conceptual framework for coordinated research responsibilities, traceable evidence and review before integration. This public account describes responsibilities and boundaries, not the underlying communication infrastructure or proprietary execution methods.
View the existing enterprise architecture Read the Orchestrator overview
9. Evidence and verification
LUMA Trust Layer v1 publishes bounded product and evidence packages with a central Master Index. The 13 August publication baseline binds source versions, manifests and applicable technical checks. The 19 October status record separately describes repository observations and product configuration.
The October record reports that seven repositories were public, the six component heads matched the August index, and the latest available workflows at those heads had succeeded. Those runs date to August apart from the AI frontend’s scheduled CodeQL run of 12 October. The record reports no open code-scanning alerts in the six components at its observation time. These are dated observations, not new scans of private production code.
Public source packages are not verified mirrors of the current applications. Reference-package exclusions around payment, signing and delivery do not describe every live product switch. Configuration readiness is distinct from an end-to-end transaction test.
The cited record does not establish completion of legal review, an independent third-party audit or the separate prospective E4 commitment-and-reveal milestone. Repository visibility does not grant rights beyond the licence stated in each repository.
10. Current and proposed stochastic benchmarks
EuroMillions is the current public proof-of-concept domain. US Powerball and Mega Millions analytics remain proposed additional stochastic environments within the 2027 planning horizon, subject to implementation and evaluation.
Additional environments could test robustness under different rules and histories. No present availability, delivery date or predictive advantage is promised.
11. Tokens, credits and NFTs are separate actions
The following describes the 19 October 2026 recorded product scope, not a permanent price list or current offer. The live application shows the applicable requirements and transaction details before approval.
- Holding-based Rooms access
- The dated release checked a finalised holding of at least 3,787.878787 LUMA. Playing did not spend LUMA, and game points had no token or cash value.
- Token checkout
- The observed public configuration enabled USDC checkout and reported LUMA delivery readiness. This documentation update did not revalidate payment finality or delivery with a new transaction.
- LUMAKeys and analysis credits
- The recorded key configuration supported independent transferable credit codes. Buying a key, receiving a redeemable code and redeeming it for analysis credits are separate actions. Holding tokens alone does not credit an AI account.
- Genesis Gold and Forge
- The 18 October release described static visual previews and varied traits. A saved off-chain image is not an on-chain NFT.
- Optional NFT minting
- The recorded configuration enabled a separate Solana Mainnet mint approval with variable SOL mint and network costs. That flow recorded no LUMA charge. Generated variations do not establish uniqueness, scarcity, rights clearance or verified collection membership.
Credits remain the accounting unit for analyses. The October observation checked the public AI entry page; authenticated features and credit redemption were not revalidated. Product configurations may change.
No equity, profit share, guaranteed yield, liquidity or exchange listing is promised. Additional token uses require their own implementation, testing and accurate disclosure.
Inspect the current Token Portal
12. Enterprise and corporate development
The current operator is Luma Quant e.U. The stated planning horizon includes stronger infrastructure and governance, future team growth and potential B2B partnerships, alongside evaluation of an Austrian corporate structure such as a FlexCo or GmbH.
These are intentions, not completed corporate changes. A choice of company form is separate from any sector-specific authorisation a future product may require.
13. Market-intelligence translation
The research question is whether selected uncertainty-analysis mechanisms can be useful with market-specific data and evaluation. The proposed mapping includes:
- Higher-order interactions among assets, sectors and factors.
- Separate views of momentum, stability, liquidity, volatility, sentiment and risk.
- Negative-space hypotheses about missing confirmation and absent participation.
- Candidate regime changes and state-dependent behaviour.
- ENMF memory of states, transitions, failures and later outcomes.
- Sequential walk-forward evaluation and appropriate historical baselines.
- Monte Carlo scenarios, tail-risk analysis and sensitivity tests.
- Decision gates that preserve uncertainty and the option not to act.
Potential inputs include macroeconomic, microstructural, capital-market, semantic, news, on-chain and off-chain information. None of these proposed mechanisms establishes a market advantage without domain-specific evidence.
Explore the complete market research mapping
14. Digital-asset regime research
Digital assets are one proposed domain within the 2027 planning horizon. The intended work examines interactions among market context, narratives, news, on-chain activity, liquidity and risk. It is not a promise of profitable trading or a statement that an automated investment product is available.
15. Cross-domain research horizon
Further research directions include medical information, anomaly mapping, AI reliability, blockchain infrastructure, quantum suitability and potential applications. Their presence in the agenda means questions are being explored, not that useful products or scientific breakthroughs have been demonstrated.
Each domain requires its own data rights, baselines, privacy, safety and expert review. Health-related research is not diagnosis, treatment advice or a substitute for qualified professionals.
16. Open research and synthetic telemetry
The public Privacy-Preserving Synthetic Stochastic Telemetry dataset supports reproducible methodological experiments while keeping operational user data and proprietary implementation separate. Synthetic data is useful for controlled tests; success on it is not a real-world predictive result.
17. Research coordination and community participation
The research environment connects data and evidence work, theory and validation, collaborative research, new questions, and engineering transfer. Independent review remains distinct from coordination. The public research overview explains these responsibilities without publishing internal assignments or infrastructure.
Shinka is presented within experimental optimisation research. Proposed changes are assessed in their experimental scope; this does not mean the production engine modifies itself automatically.
Community members can explore the AI Platform, play together in LUMA Rooms, visit the Token Portal or share questions on the official Discord . Participation does not guarantee analytical or financial outcomes.
18. Development discipline and version history
Whitepaper v1 preserves the 2023–2026 foundation thesis. The August v2 edition describes the reported architecture and expansion direction. The 19 October addendum records product and evidence updates. The 29 October editorial update brings the public research narrative into the new website structure without re-labelling old evidence as newly verified.
Research hypotheses, synthetic results, historical observations, independent reproductions and prospective evaluations remain separate evidence classes. A finding may justify further work, a revision or rejection. Algorithm changes require separate evaluation and approval.
19. Philosophy
Robustness over hype. Evidence over certainty. Transfer only after validation.
20. Scope and disclaimer
LUMA Quant’s analytical outputs are informational and experimental. They are not financial or investment advice, medical advice or diagnosis, and do not guarantee outcomes or trading performance.
This document describes architecture, dated product observations and research intentions. It does not establish regulatory approval, legal classification of a token or transaction, completion of legal review, or a completed independent audit. Current application terms, permissions and transaction details must be reviewed before approval.