MarkdripBiometric & Voice Data Notice
Version: September 8, 2026. Applies when this version is presented with the updated application.
This notice describes how Markdrip handles voice data, including data that may be classified as biometric information under applicable law.
What Markdrip creates and stores
After you enable voice memory, naming a speaker or confirming a suggested name can automatically save the following on your device:
- A voice fingerprint — a compact mathematical representation of that person’s voice, used solely to recognize the same voice in future meetings.
- A short voice clip — up to approximately ten seconds of audio, stored so you can play back and verify who the voice belongs to.
- The name you assigned — linked to the fingerprint and clip.
These items are stored together in a local database (gallery.db) in Markdrip’s application-support directory on your Mac, separate from your meeting vault.
Under laws such as the Illinois Biometric Information Privacy Act (BIPA), the EU General Data Protection Regulation (GDPR), and the California Consumer Privacy Act (CCPA/CPRA), voice fingerprints used to identify a person may constitute biometric data, biometric identifiers, or sensitive personal information. Markdrip treats them accordingly.
Why Markdrip creates this data
Voice fingerprints and clips exist for one purpose: to label who spoke across your meetings so your notes attribute the right words to the right people. They are not used for advertising, profiling, or any purpose other than speaker recognition within the Markdrip application on your device.
Your voice memory choice
The first time you choose Remember voice, a short explanation asks you to enable local voice memory. This is a one-time awareness step for the setting. Once enabled, naming speakers can save their voices automatically while Remember voices remains on.
This awareness step does not collect permission from participants. You remain responsible for informing people and obtaining any permission required to record and remember their voices. Markdrip does not send recordings or reusable voice data to external providers.
Turn Remember voices off in Settings to stop new saving; existing voices remain until you erase them. Auto-name settings control applying known names in later meetings. Automatic names are marked “Auto” and can be corrected. Review and remove saved voices in People.
Where your voice data lives
Voice fingerprints, clips, acoustic information and voice-derived interpretations always stay on this device. Markdrip does not offer external voice processing. Cloud AI permissions cover only approved ordinary transcript words, minimal speaker labels and permitted text context; they cannot authorize sharing recordings or voice data.
For optional text AI, the selected service operates under its own retention, training, regional, and billing terms. Turning that text capability off stops future sharing but cannot recall text already sent; use its deletion controls where applicable. Voice data never enters licensing or product telemetry. You remain responsible for obtaining any required consent from people whose voices you enroll.
Retention
Voice fingerprints, clips, and associated names are retained on your device until you delete them. There is no automatic expiration. You control retention entirely through the actions described below.
Uninstall caveat. Deleting (uninstalling) the Markdrip application does not automatically delete your voice data or your meeting vault. The gallery.db file and your vault remain in their respective directories on your Mac. To permanently destroy all voice data before uninstalling, use “Erase all voice data” in Settings.
Your controls
Markdrip provides the following controls, all exercised locally on your device with no network calls:
| Action | What it does |
|---|---|
| Forget this voice | Permanently deletes all fingerprints and clips for one person. The contact record is kept but unlinked from voice data. |
| Delete contact | Removes the contact file from your vault. It does not automatically delete voice data; use “Forget this voice” first if you want both removed. |
| Discard sample | Removes a single voice clip and its associated fingerprint. |
| Erase all voice data | Permanently deletes every fingerprint, clip, and voice-linked record in the gallery. Contacts are unlinked but not deleted. This action cannot be undone. |
| Turn off speaker identification | Stops future voice matching. Optionally erases all stored voice data at the same time (you are asked). |
| View and play | Browse stored voice clips per person to verify identity. |
| Rename or merge | Correct a name or merge two identities. |
| Export consent records | Export your enrollment and consent history as Markdown or JSON for your own records. |
Consent
Before enabling voice memory or saving someone’s voice, obtain any permission needed to capture and store that person’s voice data. The global awareness step is not proof of their consent.
If you record other people in jurisdictions that regulate biometric data — such as Illinois (BIPA), the European Union (GDPR), or California (CCPA/CPRA) — those laws may require the explicit written consent of each person whose voice is fingerprinted, not only your own. Obtaining that consent is your responsibility as the person operating the recording.
Markdrip keeps local voice records that you can review and export. Existing enrollment records do not establish that a participant gave permission; keep any required participant-consent records yourself.
No sale, no sharing
Voice data is never sold, licensed for another party’s use, or used for advertising. Markdrip does not send voice data to external providers and has no server-side copy of your voice data to sell or share.