Let’s connect

RESPONSIBLE DEVELOPMENT

More context.
Clearer boundaries.

Brain recordings and health-related notes deserve deliberate controls. A useful AI workflow makes its sources, limitations and decision authority visible.

A research direction, not a cleared indication

The public in-ear EEG research foundation does not establish regulatory clearance, diagnostic accuracy, equivalence to another test or treatment benefit. Device indications and clinical use require the appropriate evidence and authorization.

Evidence before interpretation

A signal summary, participant note, research paper and reviewer decision have different evidentiary roles. Preserve the distinction and show uncertainty instead of resolving it through wording.

Consent and purpose

A connected study requires an appropriate consent process, authorized data use and a defined purpose. This website collects no EEG, symptoms, recordings or participant health records.

Access and minimization

Connect only the data needed for an agreed research task. Processing, access control, retention, deletion and any model-provider terms must be defined before production use. No HIPAA or security certification is claimed.

Human review

A generated draft cannot approve a clinical interpretation, change a study record, determine treatment or replace professional judgment. Define the reviewer role and approval status.

Reproducibility

Retain recording identifiers, source intervals, analysis versions, model context and review decisions where appropriate. A useful result should be inspectable after the conversation ends.

No secondary inference

Do not turn brain recordings or diaries into covert employee monitoring, personality judgments, advertising profiles or insurance eligibility decisions. Keep use aligned to the consented research purpose.

BEFORE YOU CONNECT A DATASET

Know what this
website does—and doesn’t.

Is this a device, an AI platform, or both?

MZR’s research foundation is an in-ear EEG interface. This website develops a software layer around that foundation: signal-aware records combined with a source-grounded language-model layer for research review. It is not a claim that all displayed software is commercially available.

What does the large language model actually do?

It reads authorized summaries, diaries, annotations, protocols and approved research material. The intended tasks include cross-night synthesis, evidence retrieval, contradiction handling, documentation checks, and structured report drafting. It does not directly turn arbitrary raw EEG into a medical diagnosis.

What is different from an ordinary sleep score?

The emphasis is on reviewable context: which records support a statement, which intervals cannot be interpreted, what changed across nights, and what a qualified reviewer still needs to check. No proprietary scoring accuracy is claimed here.

Are the waveform and study examples real patient data?

No. All traces are synthetic and all demonstration records are authored. The website does not acquire EEG, access patient systems, or call a live language model.

Is regulatory clearance or clinical efficacy claimed?

No. This website does not establish FDA clearance, a cleared indication, clinical equivalence, diagnostic accuracy or treatment efficacy. Clinical claims and product availability require separate verification.

Can our team discuss a research collaboration?

Yes. Use the collaboration page to prepare a non-sensitive project brief and access the public company contact route. Define the research question, data scope, review roles and acceptance criteria before connecting records.

THE NEXT QUESTION IS WORTH EXPLORING

Let’s bring more
context to the signal.

Start a research conversation