Product preview · Example research data. Research receipts, documentation counts and workflow capabilities below illustrate the demo. They do not establish operational security, certification or scientific validation.

Trust & security

What we can show, and what we have not yet.

How Hylē keeps research checkable is documented below. Infrastructure, subprocessors, encryption, retention and model training stay unpublished until verified operator documentation exists.

Ask about security

S1 · BR-01 · Assay dataset

Assay summary

Europe PMCOpen accessCited once in R-042Supports

Result

Across the four panels, candidate B’s relative effect was 0.82, 0.73, 0.91 and 0.99, and every uncertainty bound crossed 1. S1 The organoid panel shows the largest effect (0.73).

A receipt, as Hylē shows itS1, the record behind the first claim in the example studyOpen the study

No trust details published

No live security or incident feed is connected.

Trust details could not be loaded

Nothing was changed. Try again in a moment, or check the system status.

01Research integrity

Four commitments shape how Hylē writes a result. Each is documented, and each shows up on screen as a receipt you can open.

Citations stay with the claim

Documented

Sources remain part of the result, not a list appended at the end. Select a citation to read the passage it came from.

HylēR-042 · synthesisInsufficient evidence

Candidate B has the most consistent directional signal across these cohorts, but the evidence is insufficient to select a target.

All four cohort estimates sit left of the null line at 1 (0.82, 0.73, 0.91 and 0.99), favouring B S1. The last, the Replication-labelled row, is near-null and is not an independent replication S2. Every uncertainty bound reaches past 1, and assay comparability and a real independent replication are missing S3.

The answer in the example study. S1, S2 and S3 sit inside the sentences they support.

Uncertainty stays visible

Documented

When evidence is thin or conflicting, the result says so. Claims without a retrievable source are flagged.

Primary cohortsReplication, not independentCohortno effectEffect [95% CI]Assay panel0.82 [0.64, 1.04]Organoid0.73 [0.54, 1.02]Retrospective0.91 [0.70, 1.17]Replication0.99 [0.77, 1.25]0.500.7511.25Relative effect, log scale← favours Bfavours A →Primary cohortsReplication, not independentCohortno effectAssay panelOrganoidRetrospectiveReplication0.500.7511.25Relative effect, log scale← favours Bfavours A →
Figure 1 of the same study. The dashed row is not an independent replication, so it is drawn differently.

Runs are reproducible

Documented

Each figure records the inputs, method and version that produced it. Parameter sweeps and tests run as isolated CPU jobs.

Figure 1 · T-FORESTRun record

Measure
Relative effect (ratio)
Method
random effects
Inputs
Assay panelOrganoidRetrospectiveReplication
Recorded
inputs · method · version
Example study R-042. The figure keeps its type, measure, inputs and method with it.

Sources are open indexes

Documented

Hylē searches open scholarly indexes and keeps where each record came from and how it was accessed.

Source
S1 · BR-01
Type
Assay dataset
Index
Europe PMC
Access
Open access

IndexesOpenAlexarXivEurope PMCUnpaywall

Where S1 came from and how it was accessed.

02Data & infrastructure

These terms require verified operator documentation before publication. We list them here so the gap is visible, not hidden.

Awaiting verified documentation0 of 5published

Infrastructure

Where Hylē is hosted and who operates it.

PendingAsk
Subprocessors

Which third parties process data on Hylē’s behalf.

PendingAsk
Encryption

How data is protected in transit and at rest.

PendingAsk
Retention

How long data is kept and how it is deleted.

PendingAsk
Model training

Whether and how your work is used to train models.

PendingAsk

03Policies

The terms, privacy policy and refund policy state what Hylē has confirmed today.

Privacy policyWhere a requested workflow can travel. Terms of serviceThe terms for using Hylē. Refund policyUnused purchases, used credits and failed work. System statusCurrent service health and recent incidents.

04Need a security or privacy answer?

Send your institution’s review questions

Tell us what your institution needs to review. We will route it as a security and data privacy inquiry.

Ask about security