Privacy Policy for RapidFlow
Last Updated: September 29, 2026
BinaryScript Private Limited ("Binary Script," "we," "us," or "our") makes RapidFlow, a voice typing app for Mac (the "App"). This Privacy Policy explains what the App does with your voice and text, and what, if anything, leaves your Mac.
Company Information
BinaryScript Private Limited
India
Email: support@binaryscript.com
In short: we never receive your voice or your text. There is no account, and crash reports and usage analytics are off unless you turn them on. With the default settings, speech is recognised on your Mac and your voice and text never leave it; the only thing the App fetches from us is its daily update check. Only if you add your own API key for a cloud service does the App send your data, and then it goes directly from your Mac to the service you chose.
1. What We Collect
By default, nothing about you. The App has no account, and unless you opt in (see "Crash Reports and Usage Analytics" below) it sends no analytics, crash reports, usage data or identifiers. Downloading the App from our website is an ordinary file download.
2. What the App Keeps on Your Mac
- Dictation history: your most recent dictations (up to 100), each with its text, the time, the engine used, how long it took, and the name of the app you were typing into. They are kept for 30 days by default; you can change this in Settings > General.
- Recordings for Retry: unless you turn "keep audio" off in Settings > General, the audio of your newest dictations (up to 50) is kept so you can run one again with a different engine. When off, audio isn't kept.
- Settings and custom vocabulary: stored in the App's preferences.
- API keys you add: stored in the macOS Keychain.
- Speech models: if you choose a WhisperKit engine, its model is downloaded once, on first use, and stored on your Mac.
All of this is in your user account on your Mac. Deleting the App's data, or the App, removes it.
3. Update Checks
Once a day, the App checks for a new version by fetching a small update feed from media.binaryscript.com, and it asks you before installing anything. Like any web request, the check carries your IP address and the App's name and version; it sends no information about your Mac, your settings or your dictations. You can also check by hand from the menu.
4. Crash Reports and Usage Analytics (Optional, Off by Default)
During setup, and at any time in Settings > Privacy, you can choose to help us improve the App. Both options are off unless you turn them on:
- Send crash reports (Firebase Crashlytics, Google LLC): what the App was doing when it crashed, with technical details of the crash, your Mac's model and macOS version, and the App's version.
- Share anonymous usage (Google Analytics for Firebase, Google LLC): counts and timings such as which setup steps were finished, whether a dictation worked or failed and how long it took, which engine and language were used, which kind of voice command was used, and update choices. Google Analytics also records basic device information and an app-instance identifier. Ad personalisation is turned off.
These never include your audio, your dictated text, window titles or app names. Turning an option off stops it. Google's privacy policy: https://policies.google.com/privacy.
5. Permissions and Why the App Needs Them
- Microphone: to hear you while you hold the dictation key.
- Accessibility: to type the text into the app you are using.
- Input Monitoring: to notice when you press and release the key you hold to dictate.
6. Speech Recognition
By default, speech is recognised on your Mac by Apple's on-device speech model, and your voice never leaves your Mac. The WhisperKit engines also run on your Mac.
Optional cloud engines. If you add your own API key for a cloud speech service and choose it as your engine, the audio of each dictation is sent from your Mac directly to that service to be transcribed. The services the App supports are OpenAI, Sarvam AI, Deepgram, ElevenLabs, Groq, Google Gemini and AssemblyAI. Your use of a service is under your own agreement with it and its privacy policy; we never receive the audio or the text.
7. AI Polish (Optional)
AI polish cleans up punctuation and filler words. With Apple's on-device model, the default, it runs on your Mac. If you choose Claude (Anthropic) or Google Gemini with your own API key, the text of your dictation, not the audio, is sent directly to that service to be polished.
8. Voice Commands (Optional)
If you add your own Anthropic (Claude) API key, commands you speak with the command key can be understood by Claude. For each such command, the App sends Anthropic what you said, together with the names of your running and installed apps, a few window titles of your running apps, and the names of your Shortcuts, so Claude can choose an action. Window titles can contain what is on screen, such as a document or email subject. Without a key, or with this turned off in Settings, commands are handled on your Mac.
9. Third-party Services You Choose
These are used only if you add your own key and choose them. Each has its own privacy policy:
- OpenAI: https://openai.com/policies/privacy-policy/
- Sarvam AI: https://www.sarvam.ai/privacy-policy
- Deepgram: https://deepgram.com/privacy
- ElevenLabs: https://elevenlabs.io/privacy-policy
- Groq: https://groq.com/privacy-policy/
- Google Gemini: https://policies.google.com/privacy
- AssemblyAI: https://www.assemblyai.com/legal/privacy-policy
- Anthropic (Claude): https://www.anthropic.com/legal/privacy
10. Data Sharing
We do not sell, trade or rent your information. Your audio and text reach a third party only when you choose a cloud service as described above, and they go directly from your Mac to that service. If you opt in, crash reports and usage analytics are processed by Google as described above.
11. Your Choices
- Keep the default on-device engines and your audio and text never leave your Mac.
- Turn crash reports and usage analytics on or off in Settings > Privacy.
- Turn off "keep audio" and change how long history is kept in Settings > General.
- Remove an API key in Settings to stop using that service.
- Withdraw the Microphone, Accessibility or Input Monitoring permission in System Settings > Privacy & Security.
12. Children's Privacy
The App is not directed at children. We do not knowingly collect information from children.
13. Changes to This Policy
We may update this Privacy Policy, for example when the App gains a feature. The "Last Updated" date above shows the latest version.
14. Contact Us
BinaryScript Private Limited
Email: support@binaryscript.com
We will respond within 30 days.