Private Mac notes for user researchers

User researchers and UX design strategists are the voice of the customer inside product organizations. Through in-depth qualitative interviews, unmoderated usability tests, and generative field studies, researchers uncover the emotional friction, unmet needs, and subtle mental models that dictate product adoption.

The foundational requirement of great qualitative research is genuine human connection. When a user feels comfortable, heard, and unjudged, they reveal their honest workarounds, embarrassing mistakes, and true frustrations.

However, modern AI meeting utilities actively disrupt qualitative research. Inviting an automated third-party bot into a user interview creates an immediate clinical atmosphere. Participants become defensive, hesitate to criticize prototype designs, and worry about corporate surveillance. Furthermore, uploading customer voice files to third-party cloud transcription companies creates immense GDPR and data protection liabilities.

Oyma is the dedicated research workbench for UX researchers on macOS. Operating locally on Apple Silicon, Oyma captures interview sessions silently via ScreenCaptureKit, runs speech recognition 100% on-device with Whisper, and formats qualitative insights into durable, interconnected Markdown documents stored securely in your local vault.

The rapport crisis in qualitative research: Why bots compromise data

In user research, data quality depends entirely on the psychological state of the participant:

1. The Hawthorne effect: When an automated bot enters the video call roster displaying an "AI Recording Bot" moniker, participants modify their behavior. They speak more formally, withhold criticism of difficult workflows, and try to give “correct” answers rather than authentic reactions.
2. Participant consent and GDPR liabilities: Video recordings contain biometric voice prints and personal identifiable information (PII). Transmitting this data across external cloud transcription APIs complicates data processing agreements and increases audit exposure.
3. Manual quote extraction fatigue: Reviewing multi-hour interview recordings to locate exact verbatim quotes for executive presentations takes hours of manual scrubbing.
4. Isolated research repositories: Cloud transcription tools trap user insights in closed SaaS dashboards, making it difficult to link interview evidence directly into engineering specifications or product roadmaps.

The user researcher’s qualitative knowledge topology

Oyma organizes qualitative findings into an interconnected research graph that bridges raw user quotes with product priorities:

┌────────────────────────────────────────────────────────────────────────┐
│                   UX RESEARCH REPOSITORY TOPOLOGY                      │
├────────────────────────────────────────────────────────────────────────┤
│                                                                        │
│                    [[Synthesis: Onboarding Drop-off]]                  │
│                    research/synthesis-onboarding-2026.md               │
│                                ▲                                       │
│          ┌─────────────────────┼─────────────────────┐                 │
│          │                     │                     │                 │
│          ▼                     ▼                     ▼                 │
│  [Interview: Participant A] [Interview: Participant B] [Interview: P3] │
│  "Felt lost at step 2"      "Where does my data go?" "Too many forms"  │
│          │                     │                     │                 │
│          └─────────────────────┼─────────────────────┘                 │
│                                ▼                                       │
│                  [[Affinity Theme: Trust & Privacy]]                   │
│                  insights/themes-trust-transparency.md                 │
│                                                                        │
└────────────────────────────────────────────────────────────────────────┘

By connecting interview records to thematic synthesis notes using [[wikilinks]], you build a cumulative qualitative knowledge base. When a designer or product manager asks for evidence supporting a design change, you can link directly to the authentic user quotes with zero delay.

The 4-step user research workflow

1. Silent, natural interview capture

When you launch a user interview on Zoom, Google Meet, or Whereby, Oyma identifies the active audio stream and displays an unobtrusive desktop prompt. One click activates capture via ScreenCaptureKit. No bot appears in the participant panel, preserving a natural 1-on-1 atmosphere.

2. Live observational field notes

While observing the user navigate your prototype, you can record timestamped observations and behavioral tags (e.g., #hesitation, #confusion, #delight) directly in the Live Preview editor without breaking eye contact.

3. On-device Whisper transcription

When the session concludes, local Whisper models transcribe the dialogue directly on your Mac’s Apple Silicon Neural Engine. The engine accurately captures spoken feedback across 13 languages, automatically separating questions from participant answers.

4. Qualitative affinity mapping in Markdown

The transcribed interview writes directly to your research vault as a clean .md file. Using tags and [[wikilinks]], you can organize quotes into thematic clusters and export executive debriefs in minutes.

Authentic user research session note sample

Here is a realistic qualitative interview note generated in Oyma:

---
title: "User Interview: P-104 (B2B SaaS Onboarding Usability)"
date: 2026-10-08
type: user-interview
participant_id: P-104
persona: "Mid-Market IT Administrator"
cohort: "Trial Day 3"
tags: [research, interview, onboarding, friction, sso]
---

## Participant Background
IT Administrator managing 350 seats at a logistics company. Attempting to configure team permissions and SAML single sign-on during the self-serve trial period.

## Key Usability Observations
- **Initial setup flow:** Completed organization name and team size fields in under 45 seconds without hesitation.
- **SSO configuration friction:** Paused for 3 minutes on the certificate upload screen; confused by label wording between "Entity ID" and "Audience URI".
- **Drop-off trigger:** Attempted to test login with a secondary test account; encountered generic error code `ERR_401_AUTH` without remediation steps.

## Verbatim Participant Quotes
> "I really like the minimalist layout, but when the SAML test failed, it just gave me a blank red banner. I have no idea if my certificate was expired or if the domain was wrong." *(Timestamp: 18:42)*

> "If I can't get single sign-on working within 15 minutes, our infosec policy won't let me roll this out to our department leads, no matter how good the note features are." *(Timestamp: 24:15)*

## Usability Recommendations
1. **Error state clarity:** Replace generic `ERR_401` message with diagnostic validation tips (e.g., "Check that your Identity Provider Issuer URL matches the domain").
2. **Setup guide:** Embed a collapsible step-by-step tutorial for Okta and Microsoft Entra directly in the SSO settings modal.

## Action Items
- [ ] Researcher: Tag and link P-104 quotes into [[Synthesis: Onboarding Drop-off]]
- [ ] Product: Share error banner recommendations with the auth engineering pod
- [ ] Design: Review proposed tooltip wording with content design lead

Total privacy and GDPR compliance for participant data

User researchers carry a strict ethical and legal duty to protect participant confidentiality. Sending customer voice files to third-party cloud transcription companies creates immense regulatory exposure under GDPR and CCPA.

Oyma delivers complete data custody:

• Zero cloud egress: Audio recording and transcription run 100% on-device on Apple Silicon.
• Verifiable safety: Confirm that no participant audio leaves your workstation via the transparent Egress Log.
• Permanent open files: Notes live as pure Markdown files on your Mac’s encrypted disk, fully accessible in Obsidian and text editors forever.

Conduct qualitative research that respects participant trust, captures authentic human insights, and preserves local Markdown permanence.