On-device transcription
Transcribe speech on your PC using NVIDIA Parakeet running on the CPU. The core path does not upload audio and does not require an account.
Whisperstream
Whisperstream is a private Windows dictation app that transcribes speech locally on your PC, types into the active app, and works offline after setup.
Private dictation for Windows
Whisperstream is a private dictation app for Windows that transcribes speech on your own PC instead of sending audio to a cloud service. It is built around push-to-talk recording, local speech recognition, and direct text insertion into the app or field you are already using.
The product is positioned for people who want dictated text to work across common Windows apps without an add-in or always-on background listener. Source pages also describe optional cleanup features, transcript history, and offline operation after the model is downloaded.
Core features
Push-to-talk controls
Hold a hotkey to dictate, or switch to toggle mode for hands-free recording. The default shortcut is Right Shift, and the shortcut can be remapped.
Broad app compatibility
Deliver dictated text into the active field by clipboard paste or simulated keystrokes. The source says it works in apps such as Outlook, Word, Slack, VS Code, browsers, and terminals without an add-in.
Local transcript history
Keep searchable transcript history on the device, replay saved audio, and copy past dictations again later. History is encrypted at rest and can be disabled.
Optional AI cleanup
Clean up dictation with an optional AI step that removes filler words, fixes punctuation and grammar, and applies spoken self-corrections. Users can use a local model or a configured provider.
Custom dictionary and profiles
Add names, jargon, and acronyms once so Whisperstream uses them consistently in later dictations. The source also says you can set per-app profiles and custom prompts.
Common ways to use Whisperstream
General Windows dictation
Speak into emails, documents, chat apps, and browser fields, then let Whisperstream insert the text where your cursor already is.
Developer voice input
Use a local dictation workflow for coding prompts and other terminal or editor input, including the Claude Code workflow highlighted on the site.
Offline work
Dictate when you are offline or on an unreliable connection after setup, since the core transcription path can continue without internet.
Saved dictation history
Keep a private local transcript record for later lookup, audio replay, and copy-and-paste reuse of previous dictations.
Cleanup and vocabulary control
Reduce manual cleanup by using the optional AI step and custom dictionary entries for names, jargon, acronyms, or repeated wording.
Pros and Cons
Pros
- Core transcription runs locally on the PC, with no audio upload in the main path.
- Works across many Windows apps, including Outlook, Word, Slack, VS Code, browsers, and terminals.
- Push-to-talk is the default, which avoids a constantly listening microphone state.
- Optional transcript history stays encrypted on the device and can be disabled.
- The license is a one-time purchase, and the source says one license can be used on up to 2 of your own PCs.
Cons
- The core transcription flow is limited to supported Windows 10/11 64-bit PCs.
- Some features still depend on internet access, including license validation, updates, model downloads, and any configured online enhancement or external destination.
- Application policies, security settings, or remote-session restrictions can still block or limit text delivery in specific apps.
FAQ
Is there a free trial?
Yes. Whisperstream offers a free trial, and the source says you can download it and try it without an account or credit card.
Does Whisperstream work without an internet connection?
Core transcription and text delivery can run offline after the required speech model is downloaded. Internet access may still be used for license validation, updates, model downloads, and any optional online features you enable.
What system requirements does it have?
The source lists Windows 10 or 11 on 64-bit x64 computers. It runs on CPU and does not require a GPU; 8 GB RAM is the minimum and 16 GB is recommended.
Can I use one license on more than one PC?
Yes. One license can be activated on up to 2 PCs that you own, such as a laptop and a desktop.
What languages are supported?
Whisperstream supports 39 languages. The source notes that the language is chosen during setup and that the app downloads a matching model when needed.
Quick Facts
- Category
- Productivity
- Platform
- Windows 10/11 (64-bit x64)
- Primary use
- Private speech-to-text dictation
- Company/domain
- whisperstream.io
- Pricing
- $29 once; free trial available
- License
- One license for up to 2 PCs
Whisperstream Alternatives
SpeakoFlow
SpeakoFlow is a free, open-source desktop voice-to-text app for Windows, macOS, and Linux. Dictate into any app, use Flow, cleanup, translation, and a screen-aware assistant.
Carbon Voice
Carbon Voice is an async voice messaging app for teams and individuals, with transcripts, AI catch-up, and cross-device access.
Sanota
Sanota turns spoken memories and interviews into clear written stories for personal storytelling, family history and shared memories, with guided prompts and subscriptions.
Goldfish
Goldfish is an AI memory app for macOS that helps you reply, write, summarize, and continue with context already on your Mac. It stores memory locally and writes in your tone.
Flunkey
Flunkey is a voice-first productivity layer for Windows that turns speech into text, AI outputs, and remembered context without leaving your app.
ArtDeck
ArtDeck is a visual reference board app for iPhone, iPad, and Mac. Collect mixed references on one canvas and study them with ToolBox lenses. Offline, optional iCloud sync, one-time purchase.