Privacy first.
Security always.
Scrybapp was built from the ground up with a simple promise: your voice, your data, your Mac. No cloud servers. No accounts. No data collection. Every word you speak is processed locally by Whisper AI and never leaves your device.
How Scrybapp protects your data
Most speech-to-text applications send your audio to remote servers for processing. That means your private conversations, medical notes, legal dictations, journal entries, and personal thoughts are transmitted over the internet and stored on someone else's computer. Scrybapp takes a fundamentally different approach.
On-device processing
Every single transcription happens right on your Mac. When you press the keyboard shortcut and start speaking, your audio is captured by your microphone and immediately fed into the Whisper AI model running locally on your hardware. The audio data exists only in your Mac's memory for the brief moment it takes to transcribe, and is discarded immediately after. No temporary files, no cached audio, no recordings saved anywhere. The transcribed text is inserted directly into whatever application you're using, and that's the end of the process. Your voice data's journey begins and ends on your Mac.
Zero network transmission
Scrybapp does not open any network connections for transcription purposes. You can verify this yourself: disconnect your Mac from the internet entirely and Scrybapp continues to work perfectly. There is no server to connect to, no API endpoint receiving your audio, and no background sync uploading your transcriptions. The only network connection Scrybapp ever makes is to validate your license key during initial activation, and that single check does not include any audio data, transcription text, or usage information.
No data retention
Scrybapp does not keep any record of what you've transcribed. There is no transcription history, no log files containing your spoken words, and no database of your dictations. Once the text is typed into your application, Scrybapp has no memory of it. This is by design. We believe that a speech-to-text tool should be as transparent and forgettable as your keyboard. You don't expect your keyboard to log every keystroke and send it to a server, and you shouldn't expect that from your voice typing tool either.
Complete transparency
We don't hide behind vague privacy policies or bury data collection practices in fine print. Scrybapp's privacy model is simple enough to state in one sentence: we collect nothing. No analytics, no telemetry, no crash reports, no usage statistics, no device fingerprinting, no cookies, no tracking pixels. The only piece of data Scrybapp stores is your license key, locally on your Mac, so it can verify your purchase. That's it. You can inspect every file Scrybapp creates on your system and you won't find a single byte of your voice data or transcribed text.
What data we collect
The short answer: nothing. Here is the complete, exhaustive list of data Scrybapp collects about you and your usage.
Audio recordings
None. Audio is processed in real-time memory and immediately discarded.
Transcription text
None. Text is typed directly into your active app and never stored by Scrybapp.
Personal information
None. No account, no email, no name, no phone number required.
Usage analytics
None. We don't track how often you use Scrybapp, which apps you use it with, or how many words you transcribe.
Device information
None. We don't collect your Mac model, OS version, installed apps, or hardware specs.
Crash reports
None. Scrybapp does not phone home if something goes wrong.
Location data
None. We have no interest in where you are.
Cookies or tracking
None. The Scrybapp application contains zero tracking code.
The only data stored locally
Your license key is saved in your Mac's application support folder so Scrybapp can verify your purchase. This file never leaves your device and contains no personal information beyond the key itself.
How Whisper AI works locally on your Mac
A technical look at how Scrybapp delivers accurate, fast speech-to-text without ever connecting to the internet.
What is Whisper AI?
Whisper is an automatic speech recognition (ASR) model developed by OpenAI and released as open-source software. It was trained on 680,000 hours of multilingual audio data, making it one of the most accurate speech recognition models ever created. Unlike proprietary systems like Google Speech-to-Text, Amazon Transcribe, or Apple's built-in dictation, Whisper's model weights are publicly available. This means anyone can download the model, inspect it, and run it locally without any dependency on external servers or services.
Local execution architecture
Scrybapp bundles a highly optimized version of the Whisper model that runs natively on your Mac's hardware. On Apple Silicon Macs (M1, M2, M3, M4 and later), Scrybapp leverages the Neural Engine and GPU cores for accelerated inference, delivering real-time transcription speeds. On Intel Macs, the model runs on the CPU with optimizations for x86 architecture. The entire processing pipeline works as follows:
- Audio capture:Your Mac's microphone captures audio in real-time. The raw audio is stored temporarily in RAM (never written to disk).
- Preprocessing: The audio signal is converted to a mel spectrogram, a visual representation of the sound frequencies. This happens in memory.
- Model inference:The spectrogram is fed through the Whisper neural network running on your Mac's processors. The model decodes the audio into text tokens.
- Text output: The decoded tokens are assembled into natural language text and inserted into your active application via macOS accessibility APIs.
- Cleanup: All temporary audio data and intermediate representations are freed from memory. Nothing persists.
Model sizes and accuracy
Scrybapp includes multiple Whisper model sizes so you can balance accuracy against speed based on your Mac's capabilities. Smaller models like Tiny and Base are incredibly fast and work well for quick notes and casual dictation. Larger models like Small and Medium provide higher accuracy for professional use cases like medical transcription, legal dictation, or content creation. All models run entirely on your Mac. There is no "premium cloud model" that requires an internet connection. Every model available in Scrybapp works offline, on an airplane, in a cabin in the woods, or anywhere else you take your Mac.
Multi-language support without cloud dependency
Because Whisper was trained on multilingual data, Scrybapp supports transcription in over 100 languages without needing to download language packs or connect to language-specific cloud services. Whether you're dictating in English, French, Japanese, Arabic, or any other supported language, the processing happens locally on your Mac with the same privacy guarantees. The model automatically detects the language you're speaking, so you can switch between languages mid-sentence without changing any settings.
Scrybapp vs. cloud-based alternatives
See how a local-first approach to speech-to-text compares with services that process your voice in the cloud.
| Feature | Scrybapp | Cloud Services |
|---|---|---|
| Audio processing location | 100% on your Mac | Remote servers |
| Internet required | No | Yes, always |
| Voice data transmitted | Never | Every word |
| Data retention policy | Nothing stored | Varies (often 30+ days) |
| Account required | No | Yes |
| Third-party data sharing | Impossible | Per privacy policy |
| Works on airplane | Yes | No |
| HIPAA / GDPR friendly | Inherently (no data leaves device) | Requires enterprise plan |
| Pricing model | One-time payment | Monthly subscription |
| AI model transparency | Open-source (Whisper) | Proprietary / closed |
Cloud services referenced include Google Speech-to-Text, Amazon Transcribe, Microsoft Azure Speech, and Otter.ai. Comparison based on publicly available privacy policies as of 2026.
No account required. Ever.
Most software today demands your email address before you can even try it. Then comes the verification email, the password creation, the marketing opt-in, and before you know it, you're in a database being tracked and targeted. Scrybapp rejects this model entirely.
Download Scrybapp, open it, and start transcribing. That's it. The free trial gives you 3 minutes of transcription with zero sign-up. If you decide to purchase, the transaction happens through a secure payment processor, and you receive a license key. Enter the key in Scrybapp, and you're done. We don't store your email, we don't create a user profile, and we don't build a shadow account for you behind the scenes.
This isn't just a privacy feature. It's a philosophy. Software should work for you without demanding your identity in return. Your voice typing tool doesn't need to know who you are to function, so Scrybapp doesn't ask.
Getting started takes 30 seconds
Download Scrybapp from our website
Drag it to your Applications folder
Open it and grant microphone access
Press Option + Space and start speaking
No email. No password. No sign-up form.
Open-source AI you can trust
Scrybapp is powered by OpenAI's Whisper, an open-source speech recognition model. This matters for privacy because it means the AI processing your voice is fully transparent and auditable.
Inspectable code
Whisper's source code and model architecture are publicly available on GitHub. Security researchers and privacy advocates can verify exactly how the model processes audio. There are no hidden features, no secret data collection, and no obfuscated network calls.
Community audited
Thousands of developers and researchers worldwide have examined Whisper's codebase. This community oversight provides a level of security scrutiny that no proprietary, closed-source speech recognition system can match. If there were hidden data exfiltration, someone would have found it.
No vendor lock-in
Because Whisper is open-source, Scrybapp's privacy model doesn't depend on a single company's promises. Even if Scrybapp as a company disappeared tomorrow, the underlying AI model would continue to exist and could be run by anyone, anywhere, without any cloud dependency.
Privacy FAQ
Does Scrybapp send my voice data to the cloud?
No. Scrybapp processes all audio locally on your Mac using Whisper AI. Your voice data never leaves your device. You can verify this by disconnecting from the internet and using Scrybapp normally. It works exactly the same because there is no cloud component to the transcription process.
Does Scrybapp require an account or email address?
No. Scrybapp does not require any account, email address, or personal information to use. The free trial works without sign-up. If you purchase a license, you receive a key that you enter directly in the app. We do not create user accounts or profiles.
Can Scrybapp be used for HIPAA-compliant dictation?
Because Scrybapp processes all audio locally and transmits zero data to external servers, it avoids the primary privacy concerns associated with cloud-based transcription in healthcare settings. However, HIPAA compliance involves your entire workflow, not just one tool. Consult with your compliance team about your specific use case.
Is Scrybapp GDPR compliant?
Scrybapp exceeds GDPR requirements by design. GDPR regulates how companies collect, process, and store personal data. Since Scrybapp collects no personal data whatsoever, there is nothing to regulate. You cannot have a data breach when you have no data.
What happens to my audio after transcription?
It is immediately freed from memory. Scrybapp does not save audio files, does not create temporary recordings on disk, and does not maintain any audio buffer beyond what is needed for real-time processing. Once your words are transcribed and typed, the audio data ceases to exist.
Can I use Scrybapp on a corporate network with strict security policies?
Yes. Since Scrybapp does not make network connections for transcription, it does not conflict with corporate firewalls, VPNs, or data loss prevention policies. IT administrators can verify that the application does not transmit sensitive data. The only network connection is a one-time license validation.
How does Scrybapp compare to Apple's built-in dictation for privacy?
Apple's Enhanced Dictation (which processes locally) is similar in concept, but Apple's standard dictation sends audio to Apple servers. Scrybapp always processes locally with no option to send data to any server. Additionally, Scrybapp uses the more accurate Whisper model and offers features like multi-language support and text cleanup that Apple's built-in dictation does not provide.
Does Scrybapp use analytics or telemetry?
No. Scrybapp contains zero analytics code, zero telemetry, zero crash reporting, and zero usage tracking. We do not know how many times you use the app, which languages you transcribe in, how many words you dictate, or anything else about your usage patterns.
Why privacy matters for speech-to-text
Speech-to-text technology has transformed how millions of people interact with their computers. From professionals dictating reports to students capturing lecture notes, voice typing has become an essential productivity tool. But as this technology has grown in popularity, a critical question has emerged: where does your voice data go after you speak?
The hidden cost of cloud-based transcription
Most popular speech-to-text services operate on a cloud-based model. When you speak into your microphone, your audio is recorded, compressed, and transmitted over the internet to remote servers. These servers process your audio using AI models and send the transcribed text back to your device. This round trip takes time (introducing latency) and exposes your voice data to numerous privacy risks.
Your audio passes through your internet service provider, potentially through multiple network hops, and arrives at data centers operated by large technology companies. Once there, your voice data may be stored for days, weeks, or indefinitely depending on the service's data retention policy. Some services use your audio data to train and improve their AI models, meaning your private dictations become part of a dataset that is processed by algorithms and potentially reviewed by human annotators.
Local transcription: a fundamentally better approach
Local speech-to-text processing eliminates these privacy concerns by keeping everything on your device. When you use an offline voice typing application like Scrybapp for macOS, your audio never travels over a network. It is captured by your microphone, processed by the AI model running on your Mac's own processors, and converted to text in real time. The entire process happens in your Mac's memory and takes milliseconds.
This approach offers several advantages beyond privacy. Local processing means zero dependency on internet connectivity. You can dictate documents on an airplane, in a remote cabin, or anywhere your Mac goes. There are no API rate limits, no usage quotas, and no monthly subscription fees that increase with usage. The Whisper AI model running on your Mac provides the same high-quality transcription whether you transcribe one sentence or one million words.
Who needs private speech-to-text?
While everyone deserves privacy, certain professions and use cases make local transcription essential rather than optional. Healthcare professionals dictating patient notes handle protected health information that is subject to strict regulations like HIPAA. Lawyers dictating case notes work with privileged attorney-client communications. Journalists interviewing confidential sources need assurance that their recordings are not accessible to third parties. Business executives dictating strategic memos handle market-sensitive information. Therapists and counselors taking session notes work with deeply personal patient disclosures. Writers and authors working on unpublished manuscripts need to protect their intellectual property.
For all of these professionals, a private dictation app that processes speech locally on their Mac is not a luxury. It is a professional requirement. Scrybapp was designed to meet this need without compromise, providing accurate, fast, offline speech recognition powered by Whisper AI with absolutely zero data collection.
The future of private voice technology
As AI models become more efficient and consumer hardware becomes more powerful, local processing is becoming increasingly viable for complex tasks that previously required cloud infrastructure. Scrybapp represents this future: professional-grade speech recognition that runs entirely on your personal computer without any cloud dependency. We believe that privacy should not be a premium feature or an enterprise add-on. It should be the default. That's why every copy of Scrybapp, from the free trial to the lifetime license, includes the same complete privacy protections. Your voice is yours, and it should stay that way.
3 minutes free. No account needed. See pricing