Privacy Policy

QuillMic does not collect personal data. It performs speech recognition on your device by default and only contacts external services if you explicitly configure it to do so using your own third-party account credentials. We operate no server-side analytics, telemetry, crash reporting, or advertising infrastructure. We do not have, request, or process accounts, email addresses, payment details, identifiers, device fingerprints, location data, or biometric data. The remainder of this Policy describes these practices in the formal terms expected of a binding privacy notice.

For the purposes of the EU General Data Protection Regulation (Regulation (EU) 2016/679, "GDPR"), the United Kingdom GDPR, and any comparable legislation that imposes the role of a "controller" or equivalent, QuillMic acts as the data controller only with respect to the limited correspondence you initiate with us (for example, by sending an email to the contact address listed in Section 13). For all data generated by your normal use of the Application — including audio, transcripts, vocabulary, hotkeys, and preferences — you are the sole controller. Such data remains on your device and is not transmitted to us.

We do not collect, receive, or store any of the following from users of the Application:

• account, login, or registration information;
• names, email addresses, postal addresses, or telephone numbers;
• payment, billing, or financial information;
• device identifiers, advertising identifiers, IP addresses, or MAC addresses;
• operating-system telemetry, crash reports, or diagnostic logs;
• usage analytics, feature-flag exposure data, or behavioural events;
• microphone audio, transcribed text, or any derivative thereof;
• geolocation data;
• biometric identifiers, voiceprints, or speaker-embedding vectors;
• cookies, web beacons, tracking pixels, or similar technologies (the Application does not embed a web browser surface for marketing purposes).

No QuillMic-operated server, database, or backup system stores data about you. There is no operator account associated with the Application.

When you activate dictation, audio captured from your microphone is processed locally on your Apple silicon Mac by the speech-recognition engine you have selected (for example, OpenAI Whisper, Apple Speech, or FluidAudio Parakeet, collectively the "On-Device Engines"). Audio buffers are held only in volatile memory for the duration of recognition and are discarded once the resulting transcript has been produced and inserted into the focused application.

The resulting transcript is delivered to the application that held keyboard focus at the time you invoked dictation. QuillMic does not retain a persistent log, history, or archive of your transcripts. Any retention of your text thereafter is governed exclusively by the application that received the text and by macOS itself.

On-device speech recognition models distributed with or downloaded by the Application are stored within the macOS application support directory allocated to QuillMic and are accessed solely by the Application.

We recognise that not every Mac is powerful enough to run the largest on-device speech-recognition or language models at the quality some users require. For this reason, the Application permits you — at your sole discretion — to route audio or text to one or more external services for higher-quality speech recognition, language-model polish, or assistant responses (collectively, "Optional Cloud Services"). At the time of publication of this Policy, supported Optional Cloud Services include those operated by OpenAI, L.L.C. and similar providers selected by you.

You should be in no doubt about what enabling an Optional Cloud Service means: any audio you dictate and any text the Application sends to that service will leave your computer and be transmitted to a third-party server outside our control. Once that data has left your device, its handling is governed entirely by the receiving provider's terms of service, privacy policy, retention practices, jurisdictional rules, and security posture. We have no ability to recall, redact, or audit data that you have caused to be sent in this manner.

Optional Cloud Services are disabled by default and remain disabled until you take two affirmative actions: (i) you supply your own valid API key or credentials for the relevant provider in the Application's settings, and (ii) you enable the corresponding feature for the specific workflow in question. The Application makes the operative state of every Optional Cloud Service visible to you within its interface — including, where applicable, an in-window indicator that data is about to be transmitted off-device, and clear labelling of any preset that relies on an Optional Cloud Service. We do not enable any Optional Cloud Service silently, and we do not present cloud-dependent features without identifying them as such.

When an Optional Cloud Service is active, the relevant payload (audio, text, or both) is transmitted directly from your device to the chosen provider over Transport Layer Security (TLS 1.2 or later). We do not proxy, intercept, mirror, tee, log, store, or retain a copy of that transmission, and we do not operate any infrastructure that touches it. Your contractual and statutory relationship for that processing exists solely between you and the third-party provider.

You may revoke your use of any Optional Cloud Service at any time by removing the credentials and disabling the feature in the Application's settings, after which the Application will return to local-only processing on your next dictation.

The Application requests the following macOS permissions, each of which is governed by your operating system and revocable through System Settings at any time:

• Microphone access— required to capture the audio you intend to dictate. Audio is processed as described in Section 4.
• Accessibility access — optionally requested to enable reliable in-place text insertion. Where granted, this permission is used only for the purpose of writing transcribed text into the focused field; the Application does not read, monitor, or record content that you have not explicitly dictated.
• Automation events — used only where you configure assistant commands that interact with another application; events are dispatched on your behalf and only upon your explicit request.

These permissions are evaluated and enforced by macOS. Revoking them does not delete data on our side because we hold no such data.

To the limited extent that processing of personal data may occur in connection with the Application — namely, the on-device processing of your own audio and any direct correspondence you initiate with us — the legal bases relied upon are:

• Consent(Article 6(1)(a) GDPR), where you actively enable a feature such as an Optional Cloud Service;
• Performance of a contractor pre-contractual steps (Article 6(1)(b) GDPR), where applicable to your use of the Application under its end-user licence agreement;
• Legitimate interests(Article 6(1)(f) GDPR), strictly for responding to support correspondence you initiate with us.

Depending on your jurisdiction, you may have rights to access, rectify, erase, restrict, port, or object to the processing of personal data relating to you, as well as the right to lodge a complaint with a competent supervisory authority. Because we do not collect or hold personal data, these rights have no practical subject matter on our side. Where you exercise such rights against an Optional Cloud Service provider, you should direct your request to that provider. We will reasonably cooperate with any lawful request directed to QuillMic that concerns correspondence you have initiated with us.

California residents have rights under the California Consumer Privacy Act ("CCPA") as amended by the California Privacy Rights Act. We do not "sell" or "share" personal information as those terms are defined under the CCPA, and we have not done so in the preceding twelve months.

The Application is a general-purpose productivity tool and is not directed to children under the age of thirteen (13). We do not knowingly collect personal data from children. Because we do not collect personal data from any user, no special children's data retention or deletion procedure is required. Parents or guardians who believe their child has interacted with us may contact us at the address in Section 13.

Because no personal data is transmitted to us, there are no international data transfers performed by QuillMic. If you choose to enable an Optional Cloud Service, any cross-border transfer that occurs is performed directly between your device and the third-party provider and is governed by that provider's transfer mechanisms (for example, the EU Standard Contractual Clauses, where applicable).

The privacy of the people who use QuillMic is not, for us, a compliance line item. It is the principle the Application was designed around. We believe you have an unconditional right to speak to your own computer and have what you said stay on your own computer. Everything within our power has been directed toward making that the default, the easy path, and the working configuration on every Mac the Application supports — from shipping a competent on-device speech-recognition model in the bundled installation, to keeping audio buffers in volatile memory only, to retaining no transcript history, to operating no server that touches your content.

We also recognise an honest limit. Not every Mac is powerful enough to run the highest-quality on-device models at the speed or accuracy that some users require, and some workflows genuinely benefit from a frontier cloud model. We do not consider it our place to deny that choice to the people who need it. The Application therefore offers Optional Cloud Services as described in Section 5. We want you to understand, with no ambiguity, what that offer entails: if you elect to use an Optional Cloud Service, the audio and text involved in that workflow will be transmitted from your computer to an external server operated by a third party, and from that moment its handling is no longer governed by us or by the Application. We surface that fact in the interface, we never enable it on your behalf, and we never make a cloud-dependent feature the default. Whether to take that step is, and remains, strictly your choice.

We do not assist any party — including law-enforcement bodies, advertisers, data brokers, analytics providers, employers, insurers, family members, or any third party whatsoever — in accessing your audio, your transcripts, your vocabulary, your configuration, or any derivative thereof, because the Application has been built so that we are not in possession of that information in the first place and have therefore no capability to disclose it. We have not implemented and will not implement backdoors, lawful-intercept hooks, silent capture, transcript exfiltration, or any other mechanism designed to circumvent the local-first guarantees set out in this Policy.

At the technical level, the Application is distributed exclusively through the Apple Mac App Store (or, where permitted, as a developer-signed and notarized build), runs inside the macOS application sandbox under the hardened runtime, ships with the network entitlement restricted to the minimum necessary for Optional Cloud Service use, and conducts any outbound communication over Transport Layer Security (TLS 1.2 or later). No security control is absolute; however, because we hold no personal data of yours on any system we operate, the maximum impact of any breach of our infrastructure on your personal information is, by design, nil.

We may amend this Policy to reflect changes in the Application's functionality or in applicable law. The "Effective date" and "Version" identifiers at the top of this document indicate the currency of the Policy. Material changes will be communicated within the Application or on its website prior to taking effect. Continued use of the Application after the effective date of a revised Policy constitutes acceptance of that revised Policy to the extent permitted by applicable law.

Questions or concerns regarding this Policy may be directed to the QuillMic team through the contact form on this site. The form delivers your message over Transport Layer Security to a private support channel monitored by the team. Marking your subject line "Privacy" helps us route the request quickly.

If you reside in the European Economic Area or the United Kingdom and remain dissatisfied with our response, you may lodge a complaint with your local supervisory authority.