Qui détient la garde des actifs ?

Termélaison n'agit pas comme dépositaire. Les actifs restent détenus sur l'infrastructure de garde choisie par l'investisseur ou son intermédiaire ; le moteur IA transmet des instructions d'ajustement, il ne prend jamais possession des fonds.

L'algorithme est-il consultable ou reste-t-il une boîte noire ?

Les paramètres de décision — seuils de risque, pondérations de signaux, historique d'ajustement — sont visibles dans l'interface. Le modèle lui-même n'est pas publié, mais chaque action qu'il déclenche est tracée et justifiable a posteriori.

Comment la liquidité des positions est-elle gérée ?

Le moteur évalue la profondeur de marché avant toute réallocation. Lorsque la liquidité d'un actif se réduit, l'exposition correspondante est diminuée en priorité, afin de limiter l'impact de marché lors d'un ajustement.

Termélaison — visualization of market data streams analyzed by artificial intelligence

Crypto risk management by AI

Protect your crypto exposure with continuous predictive modeling.

Termélaison monitors markets in real time and automatically adjusts your portfolio's risk level, 24 hours a day. No manual intervention required to react to volatility.

Access the analysis Consult the methodology

Continuous analysisIngestion of constant flow of market data

Autonomous adjustmentReallocation of risk without human delay

Seamless playbackTraceable and searchable algorithmic decisions

Methodology

A decision architecture, not a blind prediction

The system does not attempt to anticipate every market movement. It reduces exposure when structural risk increases, and restores it when conditions stabilize.

01 — Protection

Capital protection

Dynamic exposure thresholds limit potential loss on each position. These thresholds are not fixed: they tighten automatically when the measured volatility exceeds the usual intervals for the asset.

02 — Real-time analysis

Real-time analysis

Price, volume and liquidity flows are continuously processed by the predictive models. Each signal is weighted according to its historical reliability before influencing an allocation decision.

Market Flow → Predictive model

Risk assessment → Exposure adjustment

Execution → Logging

Simplified diagram of the processing cycle. Each step produces a trace that can be consulted a posteriori.

The AI engine

Three functions, one decision cycle

Ingestion, modeling and execution form a closed loop. No step works without the validated results of the one that precedes it.

Step 01

Data collection

Continuous aggregation of prices, order books, volumes and liquidity indicators on the main crypto exchanges. Raw data is cleaned before any processing.

Step 02

Predictive modeling

Statistical models assess the probability of short-term risk scenarios. The results are recalculated with each new batch of data, without locking in a single forecast.

Step 03

Autonomous adjustment

When a risk threshold is crossed, the exposure is automatically recalibrated. The action is recorded and remains viewable in the execution history.

Performance Philosophy

Risk-adjusted growth, not maximum return

Termélaison does not optimize for raw yield. The objective is to limit the magnitude of losses during unfavorable market phases, even if this means reducing participation during rapid increases.

This approach is aimed at investors who view capital preservation as a prerequisite for sustainable growth, rather than those seeking maximum exposure to volatility.

Dynamic diversification between assets according to their measured correlation
Exposure limits recalculated at each analysis cycle
Automatic reduction of leverage in periods of market tension

Termélaison — visual representation of the principle of volatility smoothing by risk management

Illustration of the volatility smoothing principle. Does not represent actual historical performance.

Professional interface

A readable dashboard for every engine decision

The interface displays active risk parameters, adjustment history and current positions. Nothing is hidden behind a single recommendation: each action is documented.

Complete history of exposure adjustments, timestamped
Risk parameters viewable by asset and portfolio
Export of decision logs for internal review

Communications encryption Two-factor authentication Access logging

PORTFOLIO / Overview Live stream

BTC ExposureModerate

ETH ExposureReduced

Volatility statusUnder control

Last adjustment4 min ago

Liquidity levelSufficient

Technical questions

Asset custody, algorithmic transparency, liquidity

The three points that come up most often among investors accustomed to traditional markets.

Who has custody of the assets?

Termélaison does not act as a depository. The assets remain held on the custody infrastructure chosen by the investor or his intermediary; the AI engine transmits adjustment instructions, it never takes possession of the funds.

Is the algorithm searchable or does it remain a black box?

The decision parameters — risk thresholds, signal weightings, adjustment history — are visible in the interface. The model itself is not published, but each action it triggers is traced and justifiable a posteriori.

How is position liquidity managed?

The engine evaluates market depth before any reallocation. When the liquidity of an asset is reduced, the corresponding exposure is reduced as a priority, in order to limit the market impact during an adjustment.

Review the methodology before any allocation.

Access the analysis Talk to the Termélaison team