Comment sont sécurisées les connexions aux plateformes d'échange ?

Les connexions aux plateformes de marché s'effectuent via des clés API à droits limités, sans autorisation de retrait. Les accès sont segmentés par fonction, et chaque appel est journalisé pour permettre un audit des opérations effectuées par les modèles.

Comment est organisée la conservation des actifs (custody) ?

Luntra Paycore n'assure pas la conservation directe des fonds. Les actifs restent détenus sur les infrastructures de custody choisies par le client ou par la plateforme d'exécution retenue, ce qui limite le risque de contrepartie centralisé sur un seul acteur.

Quelle est la structure de frais ?

La structure de frais dépend du périmètre d'accès (accès à l'analyse, exécution assistée, ou gestion déléguée) et du volume d'actifs sous supervision. Le détail précis est communiqué lors de l'échange avec notre équipe, avant tout engagement.

Comment le modèle réagit-il aux événements extrêmes (black swan) ?

Les événements extrêmes sont, par nature, faiblement représentés dans les données historiques. Pour cette raison, le système applique des limites d'exposition maximale indépendantes du signal prédictif, afin de contenir l'impact d'un scénario non anticipé par les modèles.

Quelles sont les contraintes de liquidité ?

Les décisions d'allocation tiennent compte de la profondeur de marché disponible sur chaque actif. Sur les actifs à liquidité réduite, la taille des positions et la vitesse d'ajustement sont volontairement limitées pour éviter un impact de marché défavorable lors de l'exécution.

Luntra Paycore — market data analysis using artificial intelligence

Data analysis & investment decision

Predictive Intelligence and Portfolio Optimization for Cryptoassets

Luntra Paycore combines predictive models and dynamic asset allocation. Each strategy is evaluated by backtesting on contrasting market phases before being proposed, with the objective of a more regular risk-adjusted return than passive holding.

Discover our performances Understanding the method

Approach based on rigorous backtesting cycles, re-evaluated before any deployment in real conditions.

Continuous monitoring of market signals on major assets.

Dynamic portfolio reallocation according to the detected regime.

Risk mapping updated with each analysis cycle.

A method structured in three stages, from data to decision

Before any allocation recommendation, each signal goes through a documented process. Here we detail the three stages of the cycle, without resorting to vague formulations.

01

Collecting market signals

Price, volume and market depth feeds are continuously ingested from multiple sources and then normalized to enable consistent comparison across assets and time periods.
02

Artificial intelligence analysis

The models look for statistical regularities and changes in market regimes. The goal is not to predict an exact price, but to estimate the probability of volatility and trend scenarios.
03

Risk-Adjusted Execution

The allocation decision includes exposure limitation rules. Volatility mitigation takes precedence over the search for maximum return when risk signals deteriorate.

Luntra Paycore — team overseeing market analysis models

Human supervision of models

Models do not operate autonomously and in isolation. A team monitors their behavior, documents gaps between expected performance and observed performance, and adjusts risk parameters when market conditions evolve sustainably.

This supervision aims to limit overfitting bias, common in purely algorithmic strategies applied to crypto-assets.

Historical data and backtesting results

The table below illustrates the type of comparison used during our internal backtesting cycles, pitting an artificial intelligence-driven strategy against a benchmark passive holding (buy-and-hold).

Observed indicator	AI strategy (backtest)	BTC Buy & Hold	ETH Buy & Hold
Behavior during a sharp decline phase	Exposure reduction triggered by RPM detection	Full exhibition maintained	Full exhibition maintained
Reallocation between asset classes	Dynamic, driven by market signals	No reallocation	No reallocation
Risk review frequency	Continuous, with each analysis cycle	Not applicable	Not applicable
Typical investment horizon	Medium term, adjusted according to the regime	Long term	Long term

Regulatory Disclaimer: Backtesting results are based on historical data and retrospective modeling; they do not prejudge future performance. Crypto-assets are volatile and carry a risk of partial or total loss of invested capital. Luntra Paycore does not provide personalized investment advice; the elements presented are exclusively informative and methodological.

Capital preservation at the center of the design

A prudent investor is not just looking for returns: they are looking to understand how the system behaves when conditions deteriorate. Here are the mechanisms involved.

Hedging algorithmsPartial hedging positions can be activated when the correlation between assets increases suddenly, a typical situation during phases of market stress.
Market regime detectionThe system distinguishes between trend, range and high volatility phases, and adapts the aggressiveness of the allocation accordingly.
Dynamic exposure limitsConcentration thresholds per asset and per risk class govern each allocation decision, independently of the signal generated.

Compliance note: Luntra Paycore operates within a documented risk management framework and transparently communicates the limitations of its models. No automated strategy can entirely eliminate the market risk specific to crypto-assets.

Continuous monitoring

The real-time monitoring system aggregates exposure by asset, recent realized volatility and the gap between target allocation and actual allocation. These indicators make it possible to verify that the portfolio remains compliant with defined risk limits at any time during the market cycle.

Reaction in the event of a market shock

During a sudden and rapid movement (flash crash), the models prioritize exposure reduction rather than finding an optimal exit point. This approach accepts giving up part of the potential rebound in order to limit the extent of the loss observed during the episode of high volatility.

Three investor profiles, three distinct uses

The same analysis infrastructure adapts to different management constraints, depending on the horizon and obligations of each profile.

Profile 01

Institutional investors

For an institutional allocation, the issue concerns the traceability of decisions and consistency with a predefined risk management mandate. The system documents each allocation adjustment and provides a searchable history, useful during periodic portfolio reviews.

Profile 02

Private wealth management

For an individual holding crypto-assets over a long horizon, the objective is often to maintain structural exposure while limiting phases of marked decline. The AI strategy then acts as a complement to an existing asset allocation, rather than as a complete replacement.

Profile 03

Business Cash Management

A corporate treasury exposed to crypto-assets must reconcile the need for liquidity and the desire for diversification. Cash availability constraints are integrated into the allocation parameters, in order to avoid positions that are difficult to liquidate quickly in the event of an operational need.

Frequently asked questions

The answers below cover the most common technical and security questions investors ask before engaging with our team.

How secure are connections to exchange platforms?

Connections to market platforms are made via API keys with limited rights, without withdrawal authorization. Accesses are segmented by function, and each call is logged to allow an audit of the operations carried out by the models.

How is the custody of assets (custody) organized?

Luntra Paycore does not provide direct custody of funds. The assets remain held on the custody infrastructures chosen by the client or by the execution platform chosen, which limits the counterparty risk centralized on a single actor.

What is the fee structure?

The fee structure depends on the scope of access (access to analysis, assisted execution, or delegated management) and the volume of assets under supervision. The precise details are communicated during the exchange with our team, before any commitment.

How does the model react to extreme events (black swan)?

Extreme events are, by nature, poorly represented in historical data. For this reason, the system applies maximum exposure limits independent of the predictive signal, in order to contain the impact of a scenario not anticipated by the models.

What are the liquidity constraints?

Allocation decisions take into account the market depth available on each asset. On assets with reduced liquidity, position size and adjustment speed are deliberately limited to avoid adverse market impact during execution.