Skip to content
Clinical continuity between sessions

See the week before the session.

Tetrahedron organizes consented between-session information, together with notes and conversation excerpts the patient explicitly chooses to send, into a short, source-linked Session Brief. Clinical judgment stays with the clinician.

Preparing for validation pilots · Clinician-led · Patient-controlled
Session Brief
Synthetic sample
Her ten days, night by night
A second short night
notably below her baseline
−1.4 SD
In the brief that dayNoteCheck-in
Supports clinical preparation. Not a diagnosis. Subject to clinical review.

The session is one hour. The context lives in the other 167 hours.

Between visits, what happened is remembered unevenly. Tetrahedron is designed to carry that context forward — consented information read against the patient’s own recent context, together with the notes and excerpts they choose to send.

1 h in session167 h between sessions
One connected workflow

Three surfaces, one thread of context.

21:08 · Menta

How did today sit with you?

HeavySteadyLighter

No rush. You can skip tonight.

Bounded by design

Short check-ins, grounding when it is wanted, and a private record the patient can review.

It is not therapy, not a crisis service, and never shows the patient a score. It works within limits the clinician sets.

The Session Brief

Select any line to see why it surfaced.

Each surfaced item is designed to show its source, why it was included, and whether clinical review is still pending.

Session BriefM. R. · 34 · 11 days since last session
Synthetic sample
1 · What changed since the last session
3 · May need attention
4 · Needs clinical confirmation
Synthetic sample data. Supports clinical preparation; does not diagnose, monitor continuously, or replace clinical judgment.
Why this surfaced
Repeats a pattern you marked earlier

Four of ten nights fell below her personal baseline, clustered Sunday to Tuesday. It is grouped here because it repeats a pattern you marked in an earlier session. A signal can have many possible causes, and it is not a clinical finding.

Source · wearable-derived · consented
Window · since last session
Baseline · her own recent context
Review · pending clinician
Patient control

The patient decides which notes and excerpts are sent.

Notes and conversation excerpts are chosen one by one. The rest of the brief — check-ins, wellbeing indicators and exercise tags — is covered by the consent agreed at enrolment, not by these switches. Below is the design for that item-level choice.

Pilot design concept · interactive · synthetic data
07:12Menta

Anything here you want to send to your therapist?

Notes and excerpts only — you choose each one

You can change this any time. No rush.

2 · What the patient chose to send2 of 3 sent
Note · 21 Jul
"Work has been loud this week."
Sent
From a conversation · 20 Jul
"I keep waking around two and cannot get back down."
Sent
Note · 23 Jul
Kept private by the patient — not visible to you.
Private
This section carries only what the patient chose to send. The rest of the brief is the consented summary. What the pilot will test is how sending, visibility and withdrawal should work in clinical practice.
A record the patient can see

The patient can review what they sent and withdraw it. If the therapist had not opened it yet, it disappears. If they had, the text is removed and they are told it was withdrawn.

Plain language only

No scores or tiers are ever shown to the patient. “A rougher night,” never “−1.4 SD.”

Boundaries

Designed around clinical boundaries.

What it is designed to do
  • Organize what changed relative to the patient's own recent context
  • Present what the patient sent as the patient sent it
  • Support clinical preparation with a consistent structure
  • Mark every interpretation as pending clinical confirmation
  • Offer bounded, non-clinical support between sessions, within clinician-defined limits
What it does not do
  • It does not diagnose, prescribe, change treatment or make autonomous clinical decisions
  • It does not detect or prevent crises, and is not an emergency service
  • It does not monitor patients continuously or promise a response time
  • It does not infer a clinical state from a wearable signal
  • It does not claim improved outcomes, saved time or reimbursement eligibility

Therapy stays with the therapist. Where the product’s reading and the clinician’s judgment differ, the clinician’s judgment is the record.

Privacy and safety

Privacy is the default state, not a feature.

Design principles for the pilot. Independent verification and formal review are part of what the pilot is for.

01

The patient chooses what is quoted

Conversation history is not shown to the therapist. A therapist sees a conversation excerpt or a note only when the patient explicitly chooses to send that item, and an unsent one is not surfaced as a hidden item.

02

Data minimization by design

The system is designed to limit information to what is needed for the experience and what the patient has agreed to share.

03

Encrypted, scoped, recorded

Encryption in transit and at rest, access scoped to the treating clinician, and access recorded.

04

Human in the loop

No clinical interpretation or treatment decision is acted on automatically. Clinician-facing insights remain pending review.

The pilot

What the pilot is designed to learn.

Five open questions. We have designed for them; we have not answered them yet.

Does a short, source-labelled brief tell a therapist anything they would not otherwise have — and is reading it worth the time it takes?

How we would look at it · clinician feedback after sessions, and a structured debrief

Preparing for product-validation pilots

Explore a pilot partnership.

We are looking for behavioral-health clinics, therapy practices and research partners to help shape a controlled product-validation pilot.

A conversation first. Scope, consent and boundaries before any deployment.
Clinician-led. Your protocols set the boundaries; we adapt the product to them.
Honest about stage. The initial pilot is designed to test usability, trust, workflow fit and operational readiness. It is not a clinical-efficacy study or a crisis-service deployment.
Built for behavioral health. Tetrahedron is ultimately designed for behavioral health and trauma-related care. Clinical evaluation in that population is a later, separate stage.

Start a conversation

Tell us who you are, the setting you work in, and what you would want a pilot to answer. We read every enquiry ourselves. We will reply with the next steps and what a potential partnership would involve.

Email us about a pilot partnership

Do not include patient-identifiable, clinical or health information in your email. We use what you send only to reply about a potential partnership. Your message goes directly to us by email — this page does not collect, store or transmit it, and there is no third-party form processor, analytics or tracking on this site.