A multimodal read of a ~40-minute dyadic conflict: vocal arousal for each partner over time, the phases it moved through, who interrupted when, and how tightly the two arousal levels tracked each other.
— min— speaking turns analyzed2 speakersAudio · language · acoustics
✦
What Attune does
Feature overview
This page is a live, de-identified preview of the analysis of one session. The full desktop app (macOS & Windows) adds capture, a session library, longitudinal tracking, reports, and privacy controls — all on-device. The panels marked Preview below are illustrative mock-ups; the downloadable app is fully functional.
Multimodal read
Vocal arousal, turn-taking, interruptions and cross-partner coupling from one recording.
In this demo
Language & voice
Communication-marker families and each speaker's vocal signature.
In this demo
Indicator network
Watch clinically-relevant indicators reorganize across the conversation.
In this demo
Progress over time
Track coupling, repair and escalation across sessions for a couple or individual.
Preview below
Reports
One-click Clinician and Couples summaries, exportable as PDF.
Preview below
Session library
Organize subjects and sessions — open, compare, or delete any of them.
Preview below
Capture & import
Record in-app or import an existing audio file — analysis runs locally.
Preview below
Privacy-first
On-device processing, encrypted archive, auto-lock, and auto-delete.
The app's toolbar: record a session or import an existing audio file, then analysis runs entirely on your device (speech recognition, speaker diarization, acoustic features). Assign each session to a subject, generate a report, and lock the archive. Raw audio is deleted right after analysis.
Session audio
Record, play back, and save the conversation for analysis. Audio is captured and stored on this device.
No recordings yet. Recording needs microphone permission and a secure (https) context or the installed app.
Vocal arousal over time
Each line is scaled to its own speaker's baseline (within-speaker z-score), so the comparison is change over time, not who was louder. Coupling is a sliding-window correlation between the two arousal series: above the line, the partners rise and fall together. Hover for values.
Headline
Near-symmetric dynamics. Talk time, turn-taking, repair attempts, and arousal are closely matched between partners, and once the conversation engaged, their arousal moved together. The pattern reads as a mutual cycle, not a one-directional one.
Safety-first by design
Attune treats individual wellbeing and risk as taking precedence over interaction metrics: where a real session shows distress or self-directed risk, that comes first. The tool is a descriptive aid for clinicians, not a risk assessment or diagnosis.
01
Structure & balance
02
Language & voice
Communication markers
Keyword-family proxies, counted per speaker — not validated clinical codes, and counts rather than severity. Both partners show the same patterns; the difference is degree.
Vocal signature
03
Indicator network over time
Exploratory
Each node is a clinically relevant indicator. An edge appears when two move together within a ~5-minute window around the current time (blue = together, amber = opposite; |r| ≥ 0.35), and node size reflects that indicator's level in the window. Press Play to watch the structure reorganize across the conversation. Exploratory: descriptive associations from a single session — illustrative, not estimated model parameters.
CN
Clinician notes
Private · clinician only
Prototype keeps notes in memory for this session only. Production should store them encrypted (e.g., Capacitor Preferences/Filesystem) behind clinician authentication.
04
Progress over time
Longitudinal · preview
Preview
When you keep multiple sessions for the same subject, Attune tracks how the interaction changes across them. Illustrative trajectory for one couple over eight sessions:
Direction of "improvement" depends on the couple's goals; arrows show change from the first session. Illustrative — synthetic trajectory, not real subjects.
05
Reports
Export · preview
Preview
For the clinician
Clinician report
Session structure & balance (talk time, turns, interruptions, repair)
Arousal timeline, phases, and cross-partner coupling
Sessions are grouped by subject (a couple or an individual). Open any session to re-view its analysis, compare across the timeline, or delete it. Everything stays on this device.
07
Privacy & data controls
Preview
Preview
On-device processing
Recording and analysis run locally; no audio or transcripts leave the machine.
● local only
Raw audio auto-deleted
The recording is discarded immediately after analysis — only the summary is kept.
● after analysis
Encrypted archive & auto-lock
The session archive is encrypted behind a clinician passphrase and locks itself after inactivity.
● locks in 5 min
Retention & deletion
Auto-delete sessions after 30 / 90 / 365 days, delete any session or subject, or wipe everything.
keep until deleted ▾
Support. If you or someone in this session is struggling or thinking about self-harm, call or text 988 (US Suicide & Crisis Lifeline, 24/7), or text HOME to 741741 (Crisis Text Line). In an emergency, call 911.
▸ Limitations & provenance
Single session. One conversation on one occasion; it cannot establish what is typical of the relationship.
Both partners had reportedly been drinking, which affects arousal, regulation, and recall.
Automated processing introduces error. Speaker labels, timing, and pitch are estimates; pitch uses a basic estimator and absolute values are approximate.
Loudness is not comparable between speakers (microphone distance unknown); only within-speaker change over time is meaningful.
Acoustic features index arousal, not specific emotions. No validated emotion classifier was used.
The coupling and network views are exploratory — descriptive associations from one session, not estimated model parameters.