Skip to main content

How it works

One evidential chain, from literature to decision

The platform runs as one sequential loop. Lumen extracts methodology consensus from the literature; Praxis turns it into a design; Veraxa validates the data; Nexora plans the analysis; Novora runs it; Oris reports.

Labtrack feeds lab outcomes back into Praxis, Sentinel audits across the chain, and Omnix and Prognos add prediction and forecasting.

The sequence

Each output is the next capability's input

Records persist between every step. As you scroll, watch Sentinel audit from above, Labtrack close the loop back to Praxis, and Omnix and Prognos branch off for prediction and forecasting.

Sentinelquality & validation audit across the chainLumenPraxisVeraxaNexoraNovoraOrisLabtracklab-in-the-loop → PraxisOmnixPrognosprediction & forecasting

Step by step

What happens at each link

  1. 01

    Lumen

    Literature retrieval and methodology consensus extraction from full-text Methods sections.

  2. 02

    Praxis

    Experimental design and statistical power computation grounded in literature evidence.

  3. 03

    Veraxa

    Data ingestion and validation against design and methodology consensus.

  4. 04

    Nexora

    Analytical workflow planning with citation-justified parameter choices.

  5. 05

    Novora

    Workflow execution with full provenance, reproducible end to end.

  6. 06

    Oris

    Research reports, decision-support briefs, and policy briefs with reproducibility certificates.

  7. 07

    Omnix

    Predictive machine learning, including antibiotic-resistance prediction, every model explainable.

  8. 08

    Sentinel

    Quality and validation audit with public benchmark reporting against versioned references.

  9. 09

    Prognos

    Biological trajectory forecasting, outbreak modelling, and antibiotic-resistance surveillance.

  10. 10

    Labtrack

    Electronic lab notebook, sample tracking, and protocol library that loops back into your analysis.

The decision-support boundary is a type, not a warning

The research-only boundary is built into the system itself, not a dialog you can click past. The platform offers no auto-execute for any clinical, policy, or institutional action.

decision_support.py
from enum import Enum
from pydantic import BaseModel

class IntendedUse(str, Enum):
    RESEARCH_ONLY = "research_only"

class AnalyticalOutput(BaseModel):
    # The decision-support boundary is a type-system primitive,
    # not an advisory guard rail. Every analytical output carries it.
    intended_use: IntendedUse = IntendedUse.RESEARCH_ONLY
    # ... evidence, provenance, citation records

Follow the chain into any single capability.