Parlo vs Wispr Flow: private dictation compared
Pick Parlo if you want on-device processing, an iPhone keyboard, and a free beta with no account. Pick Wispr Flow if you want its cloud features and platform list. Both insert text where you are typing. The table below compares platforms, processing, price, history, and export.
Wispr Flow runs on Mac, Windows, iPhone, and Android. Its privacy page states that transcription always happens in the cloud. Parlo runs on iPhone as a keyboard and on Mac as a menu-bar app. It transcribes and cleans up text on your device. Both apps insert text into whatever you were typing into.
| Parlo | Wispr Flow | |
|---|---|---|
| Platforms | iPhone, Mac | Mac, Windows, iPhone, Android |
| Where processing happens | On-device | Cloud, per its privacy page |
| Account required | No | not stated as a plain requirement |
| Price | Free during the beta | Free; Pro $15 a month ($12 a month billed yearly); Growth from $23 a month ($18 a month billed yearly) |
| Free tier | One version, free during the beta | Yes, with weekly word limits by platform |
| Dictation history | Local, encrypted; end-to-end encrypted iCloud sync on by default | A setting controls server storage of transcripts, audio, and history; default not stated |
| Export | Markdown, Mac only | not stated |
| iPhone keyboard | Yes, its own keyboard extension | not stated |
| Mac hotkey | Yes, hold Option | not stated |
| Custom dictionary | Yes, syncs to the keyboard | Yes, learns words automatically or by hand |
| Languages | English (beta) | 100 plus, with automatic detection |
What Wispr Flow does better
Wispr Flow supports more platforms today, including Windows and Android. Its language list covers over 100 languages, with automatic detection between them while you speak. Its pricing page also lists a Notetaker feature with speaker identification. An MCP integration is bundled into its free tier.
Wispr Flow's free tier also carries weekly word limits by platform. Desktop gets 2,000 words a week, iPhone gets 1,000 a week, and Android is unlimited, per its pricing page. Parlo's beta has no such cap on any platform while it remains free.
What Parlo does better
Parlo transcribes and cleans up your speech on your device. It uses Apple's speech engine and an on-device language model, so there is no cloud step in the path your words take. Parlo needs no account and costs nothing during the beta. Your local history is encrypted at rest. If you turn on sync, it moves through your private iCloud end-to-end encrypted, so Apple and Parlo cannot read it. Wispr Flow's own privacy page describes cloud transcription and a setting that controls whether it stores history on its servers; the page does not state the default.
Who should pick which
If cross-platform reach matters most to you, especially Windows or Android, Wispr Flow's platform list covers more ground than Parlo's beta does today. The same is true if you want a wide language list with automatic switching. If you want your speech to stay on your device through transcription and cleanup, plus a searchable on-device history and a free beta with no account, Parlo is built around that.
Founders and operators who dictate Slack messages, email, and notes on iPhone and Mac may prefer Parlo's on-device model. None of that text passes through a third party's cloud transcription step on its way to a keyboard or clipboard. Writers who want a searchable record of every dictated draft get that from Parlo by default. There is no cloud storage setting to weigh first, and you can turn sync off entirely in Settings. Teams already paying for Wispr Flow's cloud features across Windows or Android have less reason to switch. The same goes for anyone relying on its longer language list, since Parlo does not cover those platforms or languages yet.
Clinicians and lawyers who care about where audio and transcripts travel before they become text may also prefer Parlo's on-device path. There is no cloud transcription step to reason about. Neither app is a compliance product, so check your own requirements before relying on either one for regulated work.
People doing hands-free dictation for accessibility reasons, such as RSI, may weigh the two differently. Wispr Flow's site does not name a Mac hotkey. Parlo for Mac uses a hold-Option hotkey with a double-tap lock for continuous hands-free dictation. The iPhone keyboard's tap target sizes and VoiceOver labels were part of the build 54 accessibility pass.
Switching notes
Wispr Flow keeps your existing workflow if you already dictate through it on Windows or Android, since Parlo does not run there. If you dictate mainly on iPhone and Mac and want to try an on-device alternative, the switch looks like this:
- Install Parlo for iPhone from the public TestFlight link, or download Parlo for Mac from the beta page, or both.
- Turn on the Parlo keyboard in iPhone Settings and allow Full Access. That lets the app hand text back to the keyboard.
- Re-add names, acronyms, and terms you rely on to Parlo's custom dictionary. Your Wispr Flow vocabulary does not transfer.
- Try a normal day of dictation in both apps before deciding whether to cancel a Wispr Flow subscription. Parlo is a beta and still has rough edges.
Nothing about switching requires deleting your Wispr Flow account first. You can keep both installed while you compare them on your own vocabulary and workflow. Since Parlo is a beta, expect occasional rough edges. Reply to beta@parlokey.com with anything that looks wrong, since that feedback shapes the next build.
Frequently asked questions
Is Parlo free?
Yes, during the beta. There is no account and no subscription.
Where does Parlo process my speech?
On your device, with Apple's speech engine and an on-device language model for cleanup.
Can I try both?
Yes. Parlo installs from TestFlight on iPhone and a download on Mac. Wispr Flow has its own install pages for each platform it supports.
Which is better for accuracy?
We do not publish accuracy numbers. Parlo uses Apple's on-device engine. Try it on your own vocabulary.
Does Wispr Flow keep a copy of what I say?
Wispr Flow's privacy page describes a setting that controls whether it stores transcripts, audio, and history on its servers. The page does not state whether that setting defaults on or off.