Skip to main content
The point of doing structured clinical templates in-session is that the relevant content is already captured by the time you sit down to write the note. Rivet uses that fact: when you open a SOAP or DAP note, the right fields pull text from completed templates earlier in the same session and drop it into your draft. You can edit, append, or delete the auto-filled text — it’s a starting point, not a final answer. But the typing you don’t have to do is typing you don’t have to do.

​
What gets pulled where

​
SOAP — Subjective section

Score summaries from any screening measure the client filled out earlier in the session. The full list:
  • Depression: PHQ-9, QIDS-SR-16, CES-D-R, EPDS, PHQ-A
  • Anxiety: GAD-7, PSWQ, SPIN, OCI-R
  • Distress + trauma: K10, DASS-21, PCL-5, ITQ, PSS-10, CPSS-5, ACE
  • Functioning: WHO-5, WHODAS-12, WAI-SR
  • Eating: SCOFF, EDE-Q
  • Substance: AUDIT, AUDIT-C, DAST-10, CAGE-AID, CRAFFT
  • Attention + mood: ASRS-v1.1, SNAP-IV, MDQ
  • Sleep: ISI
  • Emotion regulation: DERS-18, ERQ, AAQ-II, MAAS, CFQ-7, SCS-SF
  • Personality + risk: BSL-23, MSI-BPD, ASQ, RFL-12
  • Support + youth: MSPSS, SDQ-self, RCADS-parent
  • Sleep diary, mood diary: narrative content
Each one renders as a one-line score summary with the severity band — e.g. PHQ-9: 14 (moderate) — and lands at the top of the Subjective field. You then add the narrative around it.

​
SOAP — Objective section

EMDR + behavioural worksheet content as narrative:
  • Target identification (negative cognition, positive cognition, VOC, SUDS baseline, body sensation)
  • SUDS check-in
  • Thought record
  • Behavioural activation log
  • Exposure log
These produce labelled-line blocks that read naturally inside a clinical note — the kind of content a practitioner would otherwise hand-copy from the worksheet into the note field.

​
SOAP — Assessment section

Clinical-reasoning support:
  • Safety plan (when the client filled the 6-step Stanley-Brown-style template earlier in the session)
  • Cognitive distortions (Distortion check with the practitioner- selected pattern)
  • CPT stuck-point log (for CPT-trained practitioners)

​
DAP — Data section

DAP collapses Subjective + Objective, so the Data section receives the combined pull — every screening score listed above plus every EMDR
  • behavioural-worksheet narrative listed above. It’s the single home for everything-the-session-produced.

​
DAP — Assessment section

Same as SOAP — safety plan, cognitive distortions, CPT stuck-point.

​
Plan and treatment-plan sections

No auto-fill. The Plan section is yours to write — interventions for next session, homework, referrals, follow-up. Same for every section of the treatment-plan template (Presenting concern, Goals, Objectives, Interventions, Measures, Review). These are formulation and decision- making fields where pre-filling would be misleading.

​
Deterministic, no AI, no inference

The auto-fill engine is a deterministic merge. When you open a SOAP note:
  1. Rivet walks the open SOAP template looking for fields marked with the rivet-autofill-from extension.
  2. For each match, it looks up the named source template (e.g. rivet-phq-9) in your session — checking completed, minimized, and active templates, in that priority order.
  3. If the source has been answered, Rivet renders it in the requested format (score-summary for screening scales, narrative for worksheets) and inserts the text into the SOAP field before the overlay opens.
  4. An “auto-filled” badge appears beside the field so you can see which sections were pre-populated.
There’s no LLM, no semantic matching, no inference about what you meant. The mapping from each clinical template to its destination field is fixed. If PHQ-9 is in the session, the PHQ-9 score-summary lands in Subjective. If it isn’t, the field is empty. That’s it.
This is the design that replaces an AI-scribe pipeline. The structured data was already captured — by the client, into the right template — so the note doesn’t need an LLM to extract it again. See Why no AI transcription.

​
What auto-fill doesn’t do

  • It doesn’t write your formulation. The Assessment section gets the raw safety-plan content if there’s a safety plan, but the clinical interpretation is yours.
  • It doesn’t generate language. There’s no paraphrasing, summarizing, or rewording. The text that lands in your field is the same text the template would produce on a clipboard copy.
  • It doesn’t overwrite your edits. Once you’ve typed in a field, the auto-filled prefix is part of your draft — re-running the auto-fill would not “refresh” your typing.
  • It doesn’t reach across sessions. Auto-fill pulls from templates completed earlier in the same video call. Last week’s PHQ-9 doesn’t carry forward — see Notes and your EHR for how that lives in your chart instead.

​
A small example

You run a 50-minute follow-up. The client fills PHQ-9 in the first 5 minutes (score 14, moderate) and a thought record in the middle 20 minutes. You open a SOAP note in the last 10 minutes. The Subjective field opens with: 
PHQ-9: 14 (moderate)
The Objective field opens with the thought-record content rendered as labelled lines: 
Situation:
What was happening?: Got an email from my boss about Monday's meeting
Your thought and feeling:
What you felt: anxious
How strong was it?: 8
The thought that went through your mind: "I'm going to be fired"
…
You type the Subjective narrative around the score, add your Objective observations above the thought-record block, write your Assessment and Plan from scratch, and the note is ready to export. Five minutes of typing instead of fifteen.

SOAP vs DAP

Picking the note format that fits your workflow.

Why no AI transcription

Why deterministic auto-fill replaces an AI scribe.

Exporting notes

Getting the finished note out of Rivet and into your EHR.