AI decision analysis platform
Marbre Rendaria aggregates multi-source market data and applies predictive models to estimate risks in real time and propose reasoned allocations, with a monitoring report sent every day.
Access the technical demonstrationMarbre Rendaria is aimed at freelancers and distributed teams who manage positions or budgets without ongoing local supervision. The issue is not the speed of execution, but the quality of the information available when deciding.
The platform does not replace the investor's judgment. It structures the available data, documents the model hypotheses and produces verifiable indicators, which can be consulted from any time zone.
Each recommendation produced by Marbre Rendaria results from a traceable sequence. No model output is presented without an indication of the process that generated it.
Market flows, macroeconomic indicators and internal time series are collected and normalized before any processing, in order to limit bias linked to a single source.
Normalized data feeds models that express their results as probabilities and confidence intervals, rather than firm predictions.
The probable scenarios are compared to a return and risk objective defined in advance, to produce an adjustable and documented allocation.
The main difficulty for mobile decision-makers is the delay between a market event and its recognition. Marbre Rendaria reduces this delay to a fixed and predictable frequency.
Three scenarios illustrate how the analysis is translated into a decision, structured according to a problem, analysis, observed result logic.
The elements below describe the integration and security characteristics of the platform, without commercial formulation.
| Category | Feature | Description |
|---|---|---|
| Latency | Flow processing | Updating indicators on a 24-hour cycle; Threshold alerts are continuously evaluated on ingested data. |
| Integration | API Connectivity | Connection via REST API for importing existing positions and exporting reports in structured format. |
| Integration | Data sources | Aggregation of market flows and public macroeconomic series, normalized before entering the models. |
| Security | Encryption | Encryption of data at rest and in transit to standards equivalent to AES-256 and TLS 1.2 or higher. |
| Models | Retraining frequency | The predictive models are re-evaluated at regular intervals in order to limit drift compared to current market conditions. |
Probabilistic models widen their confidence intervals when observed volatility exceeds historical ranges integrated during training. The daily report explicitly signals a drop in confidence level rather than maintaining a fixed recommendation, allowing the user to weight the decision accordingly.
The integration takes place via the documented REST API, which allows current positions to be imported without manual re-entry. The exported data follows a structured format compatible with common spreadsheets and monitoring tools used in portfolio management.
The report is produced automatically at a fixed deadline based on data ingested over the last 24 hours. It includes recent model accuracy, deviations from target allocation, and detected risk signals. Alert thresholds can be adjusted by the user according to their risk tolerance.
Marbre Rendaria does not promise a guaranteed result. The platform provides the analysis elements necessary to base a decision, documented and verifiable every day.