For months, Wispr Flow had been fielding the same question from Android users: When? The AI dictation startup, already established on macOS and iOS, kept demurring. Not yet, executives would say. We're still working on it.
The holdup, it turned out, had little to do with engineering bandwidth. Tanay Kothari, the company's CEO, wanted something iOS couldn't deliver—a voice experience that felt native to the entire operating system, not just bolted onto a keyboard. Android's more permissive architecture, he believed, might finally make that possible.
When Wispr Flow's Android app arrived on February 23, 2026, it brought an unusual interface: a floating bubble that appears whenever a text field comes into focus, hovering over WhatsApp chats and Gmail drafts alike. Tap it to start dictating. Press and hold for push-to-talk mode. The bubble vanishes automatically in sensitive contexts—password fields, banking apps—but otherwise follows users across the system. It's an approach that sidesteps Android's keyboard entirely, for better or worse.
"Android finally gave us the freedom to build the voice experience we always wanted," Kothari said at launch. Whether users actually want a persistent bubble trailing their thumbs remains an open question.
How It Actually Works
Wispr Flow's Android implementation requires a certain tolerance for permissions. The app needs microphone access, naturally. It also requests an accessibility service to detect text fields and insert transcriptions, display-over-other-apps privileges for the bubble itself, and battery optimization exceptions to keep everything running in the background. The company has published setup guides for Samsung, Pixel, OnePlus, and Xiaomi devices—a tacit acknowledgment that each manufacturer tends to bury these settings in slightly different places.
The privacy safeguards are more specific than you might expect. Company documentation lists 155 financial applications where the bubble won't appear, an effort to avoid recording sensitive information. The app supports Android 13 and higher, works on phones and tablets (though larger screens aren't yet optimized), and functions in most apps that use standard Android text fields. Custom input systems may break it.
In practice, Wispr Flow processes dictation in real time, adding punctuation automatically and stripping filler words. Say "4 pm, actually 3 pm" and the system updates accordingly. Numbered lists appear without manual markup. The interface offers two input modes: continuous dictation or short bursts, depending on how you prefer to work.
The Android release also brought support for more than 100 languages, including what Wispr describes as the first Hinglish model in the category to transcribe mixed Hindi-English speech in Latin script. A recent infrastructure rewrite reportedly delivers 30% faster dictation across all platforms, though the company declined to share specific latency benchmarks.
What's Still Missing
Desktop users have grown accustomed to certain features that Android customers won't find yet. Dictionary customization, text shortcuts (called Snippets), and formatting presets (Styles) all show "Coming soon" placeholders in the mobile app. Wispr says these will arrive eventually but hasn't committed to a timeline. Offline transcription also isn't supported—the app requires an internet connection, which may limit its appeal in regions with spotty connectivity.
Perhaps more consequential: Wispr Flow's Android release launched with unlimited free dictation during the launch period. The company hasn't specified when that promotion ends, which makes it difficult to assess the value proposition once standard pricing kicks in. (The usual structure includes a free Basic tier with weekly word caps, a Pro plan at $15 per month or $12 if billed annually, and custom Enterprise pricing.)
The Enterprise Angle
That Enterprise tier matters more than the consumer features, at least judging by Wispr's pitch. The company touts HIPAA compliance via business associate agreements, SOC 2 Type II certification, and ISO 27001 compliance—table stakes for selling into healthcare or financial services. A Privacy Mode toggle offers zero data retention, storing transcriptions only on-device, though it's unclear how many customers actually enable it.
By late last year, Wispr Flow claimed it had reached 270 companies in the Fortune 500 and was signing roughly 125 enterprise customers per week, growing at 40% month-over-month. Those figures predate the Android launch, so it's hard to know whether mobile access accelerated momentum or simply filled a gap enterprises were already asking about.
Founded in 2021 by Kothari and CTO Sahaj Garg, Wispr AI had raised $81 million as of November 2025. Its most recent round—$25 million led by Notable Capital—reportedly valued the company at $700 million post-money, a figure that reflects investor enthusiasm for workplace productivity tools more than proven revenue scale.
A Crowded Field, Sort Of
AI-powered dictation apps remain oddly scarce on Android, especially compared to the glut of task managers and note-taking apps. Google's built-in Gboard voice typing covers the basics. Typeless, which shipped its Android keyboard earlier this year, offers AI transcription in a more conventional form factor. Beyond that, the category feels underpopulated—perhaps because accurate speech recognition has become commoditized, or perhaps because most users don't dictate enough text to justify a standalone app.
Wispr Flow's floating bubble distinguishes it from keyboard-based competitors, though whether that's an advantage or an annoyance depends largely on how you work. iOS forced the company into a keyboard when it launched on that platform, a constraint of Apple's more restrictive permission model. Android allowed something more intrusive, and potentially more useful. The trade-off is that users must grant Wispr Flow extraordinary access to their devices—the kind of permissions that typically require trust, or at least a willingness to experiment.
The app is available now, distributed via a download link on the company's website. (A direct Google Play Store listing wasn't immediately visible in search results at launch, though that may have changed since.) For now, Wispr Flow seems to be betting that professionals tired of thumb-typing will tolerate a floating bubble in exchange for system-wide voice input.
Whether they actually will—well, that's the experiment.
