Saylark Privacy Policy

Last updated: 2026-04-07 Effective: 2026-04-07

1. Introduction

Saylark is a macOS voice dictation and meeting transcription app operated by Logic Fusion, LLC ("Logic Fusion", "we", "us", "our"), a Delaware limited liability company. This policy explains, in plain English, what data we handle, why, and what rights you have.

If you have questions, email voice@logicfusion.net.

2. What We Collect

We try to collect as little as possible. Here is the full list:

Audio you record for dictation. Captured locally. Processed locally. See Section 4.
Audio you record for meetings. Captured locally, stored as WAV files in your Saylark application support folder on your Mac.
Transcribed text. Stored locally. For AI features, text is transmitted (not the audio) to OpenAI through our secure proxy.
Account information (only if you sign in for team features): email address, display name, optional avatar image.
Team data (only if you use teams): team membership, shared tasks, shared dictations, usage statistics, name links, invitations.
Payment information (only if you subscribe): processed entirely by Stripe; we never see or store your card number.
Device metadata sent by Apple (through the Mac App Store, if applicable): device model, OS version, crash logs. We do not correlate these to you.
Voice embeddings (256-dimension vectors computed from speaker audio), ONLY IF the Voice Library feature is enabled. It is currently disabled by default.

3. What We Do NOT Collect

We do not collect your raw dictation audio on our servers.
We do not collect your keystrokes, clipboard contents, or text from other applications beyond what you explicitly dictate or transcribe.
We do not fingerprint your device for advertising.
We do not sell your data. Ever.
We do not share your data with advertising networks.
We do not train machine-learning models on your content.

4. How We Use Your Data (By Feature)

Local dictation (default)

Audio is captured, transcribed on-device using WhisperKit or NVIDIA Parakeet, and then discarded. The resulting text is stored in your local SwiftData database. No audio leaves your Mac.

AI features (Max mode, meeting summaries, speaker identification, style processing, daily summaries, text-to-speech)

The transcribed TEXT is sent to OpenAI through a Supabase Edge Function we operate. Our OpenAI API key is stored server-side, never in the app binary. Per the OpenAI API Data Processing Addendum, OpenAI does not use API data to train its models.

Team features (optional)

If you create an account, data you explicitly share with teammates (tasks, meeting items, stats, dictation links) is stored in our Supabase database.

Meeting recording

Audio is recorded locally and stored on your Mac. If you are on the Pro plan, the transcript (not the audio) may be sent to OpenAI for summary, action-item extraction, and speaker identification. You control this per recording.

Payments

Stripe handles all card data. We only receive a token and subscription status.

Voice Library (currently disabled)

If enabled in the future, a mathematical "voiceprint" (256-dim embedding) is computed from each speaker and stored locally. This may be considered biometric data under laws like BIPA. See Section 11.

5. Sub-processors

OpenAI, LLC — AI text generation and TTS. Data: transcribed text. US. https://openai.com/enterprise-privacy
Supabase Inc. — Database, authentication, storage, edge functions. Data: account info, team data, shared items, optional voice embeddings. US. https://supabase.com/security
Apple Inc. — macOS frameworks (ScreenCaptureKit, EventKit, etc.), Mac App Store delivery. Device-level; we do not share app data with Apple.
Stripe, Inc. — Subscription payment processing. Data: payment tokens, billing email. US. https://stripe.com/privacy

For the always-up-to-date list, see SUBPROCESSORS.md.

6. Legal Basis for Processing (GDPR)

Performance of a contract — delivering the dictation, meeting, and team features you asked for.
Legitimate interest — securing our service, preventing abuse, and improving reliability.
Consent — first-run acceptance of this policy; any optional telemetry; meeting recording.
Legal obligation — when required by law.

7. Data Retention

Local dictation transcripts: kept on your Mac until you delete them.
Meeting audio and transcripts: kept on your Mac until you delete them.
Account data in Supabase: retained while your account is active. Deleted within 30 days of account deletion.
Team data: retained while the team exists. Deleted within 30 days of team deletion.
Stripe billing records: retained per Stripe's policy and applicable tax law (typically 7 years).
Server logs (Edge Function): maximum 14 days, then purged.

You can delete everything at any time from Settings → Privacy & Legal → Delete My Data.

8. International Transfers

Our sub-processors operate primarily from the United States. If you are in the European Economic Area, the United Kingdom, or Switzerland, your data may be transferred to the US. We rely on the European Commission's Standard Contractual Clauses (SCCs) and each sub-processor's certifications.

9. Your Rights

You have the right to:

Access your data — use Settings → Privacy & Legal → Export My Data to get a JSON file of everything.
Delete your data — Settings → Privacy & Legal → Delete My Data.
Port your data — the export file is in standard JSON.
Correct inaccurate data — edit in-app or email us.
Object to or restrict processing — email voice@logicfusion.net.
Withdraw consent at any time — by deleting the app or your account.
Lodge a complaint with your local data-protection authority.

We respond to requests within 30 days.

10. Meeting Recording — Your Responsibility

Saylark can record audio from your meetings, including other participants and system audio. Many jurisdictions — including California, Illinois, Florida, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania, and Washington in the United States, and most of Europe — require consent from ALL parties before recording a conversation. Recording without consent may be a crime.

You are responsible for obtaining that consent. Saylark shows a disclaimer before your first recording and periodically thereafter. We do not provide legal advice.

11. Voice Biometrics (BIPA)

The Voice Library feature, if enabled, stores a numerical speaker embedding on your Mac. Under the Illinois Biometric Information Privacy Act ("BIPA") and similar laws, this may be considered biometric data. Embeddings are:

Generated and stored locally. Never uploaded unless you explicitly share a meeting.
Used only to cluster and label speakers in YOUR meetings.
Deletable at any time by removing the meeting.
Currently disabled by default.

If you are an Illinois resident, by enabling Voice Library you consent to local storage of voice embeddings for the purposes described above.

12. Children

Saylark is not directed to children under 13. We do not knowingly collect data from anyone under 13. If we learn we have, we delete it. First-run consent requires users to confirm they are at least 13.

13. Security

On-device data is protected by macOS file system protections and, where possible, the user's keychain.
Traffic to our Edge Function and Supabase uses TLS 1.2+.
Supabase enforces row-level security so team members can only see their own team's data.
Payment data is handled by Stripe (PCI-DSS Level 1).
We review access to production systems and rotate credentials regularly.

No system is perfectly secure. If you believe you have found a vulnerability, email voice@logicfusion.net.

14. California Residents (CCPA / CPRA)

California residents have the right to know, delete, correct, and opt out of "sale" or "sharing" of their personal information. We do not sell or share personal information for cross-context behavioral advertising. To exercise your rights, use the in-app Export / Delete buttons or email voice@logicfusion.net. We will not discriminate against you for exercising your rights.

Categories of personal information collected in the last 12 months: identifiers (email), commercial information (subscription status), audio/voice data (local only), internet activity (feature usage), and inferences (speaker labels). See Section 2.

15. Changes to This Policy

We will notify you in-app of material changes and update the "Last updated" date. Continued use after a change means you accept it. Previous versions are available on request.

16. Contact

Privacy: voice@logicfusion.net
Security: voice@logicfusion.net
General: voice@logicfusion.net