Handy vs Superwhisper: Free Local Dictation or Custom Modes?
Compare Handy and Superwhisper on offline setup, free and paid plans, platforms, custom modes and the work needed to get usable text.
Updated
TL;DR
Handy is the starting point if you want a free, open-source desktop dictation tool and are willing to choose a local speech model. Superwhisper deserves the closer look when configurable voice and AI modes, mobile use or a commercial support path matter more than avoiding a license cost.
Both belong on an offline dictation shortlist. Neither should win on assumed accuracy: the model, microphone, language and amount of editing required determine the useful result.
Published by Utter, a competing dictation app. This is an editorial comparison of official documentation, not a hands-on benchmark or an independent review.
What are Handy and Superwhisper?
Handy turns speech into text at the cursor, using a shortcut and local recognition. Its open-source repository covers desktop releases and setup. Superwhisper combines dictation with configurable modes for how speech is transcribed and rewritten.
That difference suggests a useful question: do you want the words you said, or a workflow that transforms those words into a particular kind of output?
| Decision | Handy | Superwhisper |
|---|---|---|
| Primary appeal | Free software with inspectable source | Configurable voice and AI modes |
| Desktop scope | Mac, Windows and Linux | Mac and Windows |
| Mobile requirement | Do not assume a mobile app from desktop availability | iPhone and iPad listed in Pro documentation |
| Offline setup | Download a suitable local model | Choose local processing for every active stage |
| Buying model | Free open-source project | Free tier and paid Pro licenses |
Platform and license information checked August 31, 2026 against the Handy repository and Superwhisper licensing documentation. Check the installed release for feature parity.
Handy: a useful free baseline for ordinary text entry
Start with Handy if the task is “speak a paragraph into this text field.” Avoid judging it by features you do not need. If your normal output needs only a punctuation correction, an elaborate AI workflow might add setup without saving effort.
The tradeoff is ownership of the setup. You need a suitable model, working microphone permissions and a shortcut that does not conflict with your editor. Source access is valuable if you want to inspect or change behavior, but you do not need to modify code just to evaluate the app.
For a short offline test, download the model first, disconnect networking and dictate into your normal notes app. Repeat in a browser form. Treat any optional connected processing as a separate feature and verify its data route instead of assuming the whole configuration is local.
Superwhisper: compare modes, not just the microphone button
Superwhisper’s mode documentation separates voice processing from AI processing. A mode can therefore be evaluated as a small workflow: which recognizer hears the words, what rewriting follows, and where the result is inserted.
This is useful when the same rough speech needs different outputs. A support reply may need a polished paragraph; an engineering note may need literal identifiers preserved. The test is whether the chosen mode saves editing while preserving meaning. A rewrite that sounds smooth but changes a commitment is a failure.
More controls also mean more configuration to understand. A local voice model followed by a cloud language model is not an end-to-end offline workflow.
Free versus Pro: check the exact models before paying
Handy is free. Superwhisper’s official Pro page lists $8.49/month, $84.99/year or $249.99 once, checked August 31, 2026. The purchase channel can affect price and billing support. Superwhisper Pro.
There is an important documentation mismatch: the Superwhisper homepage advertises free small models, while the Pro documentation marks local voice models as paid. Do not conclude that every local model requires Pro, or assume a particular model is free. Open the current app and check the models and modes available after the trial.
Compare the plan you actually need. A lifetime license is attractive only if the product fits and you accept the vendor’s ongoing-service terms; it is not evidence of better recognition.
How to compare accuracy and correction effort yourself
Use the same microphone and the same intended paragraph in both apps. Try this deliberately awkward input:
Send the draft to Maya, not Maria. The release is postponed until Friday. Keep the feature flag off, and preserve the identifier customer_id exactly.
Check the negation, name, date and identifier before judging punctuation. Then repeat with a natural false start and with background noise similar to your real workspace.
| Check | Record in your notes |
|---|---|
| Time until usable text | Include the editing, not only first-word latency |
| Meaning changes | Names, negations, commitments and technical terms |
| Insertion | Correct field, cursor position and line breaks |
| Offline behavior | What still works after models are downloaded |
| Recovery | Can you recover from a failed insertion without repeating everything? |
These are suggested reader tests, not measurements conducted for this article. The longer dictation evaluation guide explains how to score them.
Who should choose which?
Choose Handy first for a free desktop baseline, Linux requirements or source-level control. Choose Superwhisper when a particular mode, platform or paid feature passes your workflow test. Keep both out of your final shortlist if neither handles the language or applications you actually use.
If your requirement is specifically Mac and iPhone dictation with explicit local and BYOK routes plus reusable transcripts, Utter’s setup guide offers another path. Compare Utter vs Handy or Utter vs Superwhisper without assuming Utter is preferable just because it publishes this page.
FAQ
Does offline mean no setup or downloads?
No. A local model generally must be downloaded before disconnected use. Check the complete processing route, including any rewriting provider and storage settings.
Is open-source dictation automatically more accurate?
No. Source availability helps inspection and modification. It does not establish recognition quality for your accent or vocabulary. The Whisper project documents model and language considerations; a shared model family is not a controlled app comparison.
Should a Mac user try built-in Dictation first?
Yes, it is a useful baseline before configuring another app. Apple’s Dictation guide explains setup and language-dependent behavior. Compare the editing effort rather than assuming a paid app must be better.
Methodology and sources
Official product, repository and licensing pages were reviewed August 31, 2026. The recommendations are editorial inferences from documented workflow differences. No private customer data, hands-on testing, uptime claims or accuracy rankings are presented. For a broader shortlist, see Best Dictation Software 2026.
Published by Utter, the product discussed on this site. Comparisons reflect our perspective as its maker, not an independent testing organization. Check linked product documentation for current pricing and availability. Send a correction.
