Liardine consolidates the order books of several places, normalizes flows in real time and applies predictive models to prioritize information. You keep the decision. The platform reduces noise and decision latency.
Without custody of funds. Connection by API keys with limited permissions. Test environment available.
Switching from one interface to another fragments attention and delays interpretation. Liardine replaces this mosaic with a single, documented and controllable pipeline.
What teams see when operating without an aggregation layer.
A continuous processing chain, from raw data to calibrated recommendation.
Each step is inspectable. No transformation is hidden from the user.
Liardine is aimed at day traders and desks who already arbitrage between several sources. The platform does not promise returns. It structures the analysis to make each decision more readable and more reversible.
The parameters remain exposed: thresholds, horizons, risk tolerance. You can audit what led to a recommendation and adjust the logic to your profile.
Each module responds to a specific function. Together, they form a complete observation post, from the raw log to operational triggering.
Consolidate dispersed cash into a comparable view. Rating differences and depth become readable without changing tabs.
The models qualify the current regime and estimate the probability of persistence, with caution displayed in times of stress.
User-defined exposure thresholds, concentration alerts and step-down rules.
Transform a validated condition into a traceable action. Triggers combine signal, liquidity and risk before sending them to your tools.
Credibility relies on the transparency of the processing. Here is the sequence applied to each analysis cycle, available in the documentation.
Collection of quotation and book feeds, harmonization of formats, price steps and timestamps. Inconsistent values are isolated.
Construction of liquidity, dispersion and tension indicators. Each flow receives a reliability score before feeding the models.
Comparison of several plausible trajectories with probabilistic weighting. The dominant regime and alternative regimes are displayed.
Recommendation accompanied by a confidence level and a risk-adjusted size. The user validates or rejects, with traceability.
Each scenario follows the same grid: what you provide, what the platform processes, what you get for taking action.
When two markets diverge on the same asset, the gap may reflect a lack of liquidity rather than an opportunity.
In the stress phase, the priority becomes the containment of exposure before the search for performance.
Distorting correlations change the actual risk of a seemingly diversified portfolio.
Designed to limit rights to what is strictly necessary and keep your operations auditable.
Use keys limited to reading balances, positions and market data. Any order permission remains optional, revocable at any time from your exchange.
Your data is only used to produce your analyses. They are neither resold nor shared between customer areas. You can request the export or deletion of your history.
Latency depends on your seats, your connection and the monitored perimeter. The platform displays the standardized timestamp and the freshness status of each flow to avoid any interpretation on outdated data.
No. Liardine is a synthesis and decision support engine. It prioritizes, quantifies the uncertainty and proposes a calibration. The final order remains validated by you or by your explicit rules.
Yes. A simulation environment allows you to replay configurations, test triggers and review the methodology before any connection.
