01 / Research & analysis
AI Platform
The research-facing application for exploring LUMA’s current analysis framework and experimental benchmark.
Open AI PlatformTechnology & roadmap
Built under
uncertainty.
An analytical framework, a growing research environment and distinct ways to explore it. Connected by a common purpose: examine patterns, test assumptions and learn from what the evidence actually supports.
Development status is based on the operator’s published record. The source-code reference baseline is . These are dated records, not a live system-health report.
The existing foundation
One architecture.
Distinct responsibilities.
The public architecture connects access, application interfaces, services, analysis, memory and data. It describes the established SaaS foundation; the expanding research organisation is shown separately.
The analytical core
Several perspectives.
No unquestioned answer.
LUMA Quant examines candidate structures across analytical views of stability, change, interaction, proximity, negative evidence and ranking. Agreement is considered alongside disagreement, uncertainty and later outcomes.
A regime is a possible state of system behaviour. Regime detection asks whether relationships are changing; it does not assume every apparent shift reflects a genuine signal.
Historical evaluation, sequential walk-forward tests, simulation and post-outcome review help assess a method’s limits. The existence of a workflow does not, by itself, establish predictive value.
Explore the research questionsMemory & context
Experience needs
more than storage.
The Emergent Neuronal Memory Framework (ENMF) organises observations, relationships and analytical state. The Experience Layer is the intended connection between research findings, model-run experience and future development.
Preserve the conditions, not just the conclusion.
A useful experience record needs its source, context, version, uncertainty and limits. Negative findings and contradicted assumptions belong alongside positive observations.
The Belief State represents the current analytical state, not certainty. Proposed connections between research memory, model experience and ENMF must be evaluated and versioned before they can influence the engine.
Existing memory roles and planned integrations are distinct. This page does not assert that every Experience Layer connection is implemented or continuously learning.
Explore the existing public ENMF layer overview
- Sensory Input
- Source observations, provenance and analytical context.
- Temporal Memory
- Time windows, historical context and changing state.
- Change Memory
- Transitions and differences between observations.
- Interaction Memory
- Relationships, conflicts and cross-view behaviour.
- Negative-Space Memory
- Failures, exclusions and structures that did not persist.
- Number Landscape
- Entity-level context in the current stochastic benchmark.
- Ranking Memory
- Candidate ordering and agreement between analytical views.
- Decision Memory
- Recorded gates, selections and decision context.
- Post-Draw Learning
- The benchmark-specific name for post-outcome evaluation.
- Belief State
- The consolidated analytical state and its uncertainty.
These are public architectural roles. Scoring formulas, model weights, thresholds and execution details are not disclosed.
Research to development
How research is coordinated
A documented finding can become a proposal for engineering work. It is not an automatic instruction to change a production model.
The experimental proof of concept
A demanding test
of uncertainty.
EuroMillions is used as a deliberately severe stochastic test field. Its observable outcome history supports comparison between earlier analytical decisions and later results, while offering very little explanatory information about the next outcome.
The objective is to study behaviour under uncertainty: unstable patterns, disagreement, possible regime changes, negative evidence and failure analysis. This benchmark is a testing context, not the identity or final scope of LUMA Quant.
The boundary is explicit.
Explore the AI Platform
A synthetic demonstration is not evidence of real-world prediction. A historical observation is not a prospective result. No independently established prospective EuroMillions prediction advantage is claimed here.
Development roadmap
A foundation to build on.
A direction to test.
P0–P2 are recorded by the operator as implemented foundations. P3 combines existing product releases with continuing integration and refinement. These labels describe development, not independent assurance.
- P0 · Recorded foundation
Enterprise Product Foundation
Frontend and backend service boundaries, user access, delivery and operational observability for the first SaaS implementation.
- P1 · Recorded foundation
Applied Intelligence & Public Benchmark
Multi-axis analysis, candidate regime detection, transition and disagreement tracking, and the public EuroMillions proof of concept.
- P2 · Recorded foundation
Enterprise Data & ENMF
Semantic memory, data lineage, ETL and separation between historical information and target evaluation.
- P3 · Ongoing development
Platform & Community Utility
AI Platform, Token Portal and LUMA Rooms releases provide the basis for product improvement, feedback and carefully scoped new uses.
Source: 19 October product-status record . Roadmap intentions are not promises of delivery dates or future token rights.
The platform family
Explore what interests you.
Research, cooperative play and token participation have different purposes and their own interfaces. One does not imply the capabilities or assurances of another.
02 / Cooperative play
Playground
LUMA Rooms brings together quizzes, word associations, clues and cooperative experiences. Check access conditions in the application.
Open Playground03 / Utility & artifacts
Token & NFTs
The LUMA token, Genesis Gold, Forge and LUMAKey credit codes. Purchases, mints and code redemption remain separate user-approved actions.
Open Token PortalProduct descriptions follow the dated public product report . Current offers, prices and availability are shown by each application. The October AI check covered the public entry page, not authenticated credit redemption.
Explore the ecosystemResearch horizons
New domains.
New evidence requirements.
Methods may be investigated in other settings, but successful transfer must be demonstrated within each domain. The 2027 horizon describes an intention, not a release commitment.
Market & Digital-Asset Intelligence
Research into interactions, regime change, missing confirmation, market-state memory and scenario evaluation. Potential inputs include macroeconomic, microstructural, capital-market, semantic, news, on-chain and off-chain evidence.
Explore the market research directionAdditional stochastic benchmarks
US Powerball and Mega Millions remain proposed additional environments for testing the framework under different rule systems and histories. They are not presented as currently available products.
Medical & Complex Information Research
Exploratory work on information relationships and anomaly mapping. Any health-related application requires its own data permissions, validation, safeguards and expert review; no diagnostic capability is claimed.
Technology & New Applications
Research directions also include AI reliability, blockchain infrastructure, quantum suitability and potential applications. Their inclusion in the research agenda is not a statement of product readiness.
See the broader research agendaResponsible growth
Build capacity.
Keep the boundaries clear.
LUMA Quant’s planning horizon includes further infrastructure development, operational resilience, prospective partnerships and team growth. The operator has also described a possible future transition from Luma Quant e.U. to an Austrian corporate structure such as a FlexCo or GmbH.
These remain development intentions. Any future structure, delivery commitment or sector-specific authorisation must be established separately.
About LUMA QuantEvidence & participation
Inspect the work.
Question the assumptions.
The Trust Center connects dated product observations with the separate August public-source baseline. Repository checks do not establish an independent production audit, legal approval or completion of the prospective E4 milestone.
Public synthetic telemetry supports reproducible experimentation without exposing operational user data. Synthetic results remain separate from real-world predictive evidence.
Questions and constructive feedback are welcome. Join the LUMA community