Web Speech API vs cloud STT

Privacy, latency, accuracy, and cost trade-offs — and why VoiceMyThoughts runs on-device.

Updated April 25, 2026

Which speech engine does VoiceMyThoughts use?

VoiceMyThoughts uses Chrome's on-device Web Speech API (webkitSpeechRecognition). Audio is captured by the browser, transcribed inside the browser, and the resulting text is inserted into the focused field. The extension does not send audio to any cloud speech-to-text API.

Why this matters

Cloud speech-to-text services (the kind you call from a backend over HTTPS) are powerful but come with three trade-offs: privacy (your audio leaves your device), latency (network round-trips), and cost (per-minute pricing that limits free use). On-device recognition removes all three.

Side-by-side comparison

On-device Web Speech APICloud STT API
Audio leaves deviceNoYes
LatencyLocal — sub-secondNetwork round-trip
CostFree (built into browser)Per-minute pricing
Accuracy on common speechExcellentExcellent
Accuracy on heavy accents / specialized vocabularyGoodSlightly better
Browser supportChrome + ChromiumAny client that can call HTTP
Privacy disclosures requiredMinimalFull data-flow disclosure

When cloud STT does win

Why we picked on-device

VoiceMyThoughts is a browser extension that runs while you browse. Sending your audio to a remote API for every email, chat, and prompt would be a privacy regression for almost zero accuracy gain in everyday use. By staying on-device, the extension can deliver instant transcription, full privacy, and a Free plan that costs us nothing per minute and costs you nothing.

What this means for users

Try VoiceMyThoughts free

Add it to Chrome, click the mic icon next to any text field, and start dictating in under a minute.

Add to Chrome — It's Free