70 milliseconds. That's how long it takes for your spoken words to become text. No cloud round-trip, no buffering, no waiting. Just speak and it's there.
Need the status update before the standup — can you send it quick?
Done. Sprint velocity is up 15%, we closed 3 blockers yesterday, and the new auth flow ships Friday.
Three steps. No setup. No account.
Press and hold the key you choose (default: Right Option).
Talk at normal speed. The speech model runs locally on your Mac.
Your words are transcribed and pasted into your active app before you even move your hand.
Speed
Cloud-based dictation adds network latency, server processing time, and response delays. Just Parley runs entirely on your Mac — so your words appear the moment you stop speaking.
Cloud dictation tools make a round trip to a server. That's 200-800ms of latency before you even get a result. Just Parley processes locally in 70ms. The difference is night and day.
Cloud services throttle, queue, or disconnect during peak hours. Just Parley runs on your Mac's CPU — consistent speed, every time, no matter what.
every time, guaranteed
The moment transcription finishes, text is pasted into your active app. No confirmation dialog, no intermediate window. Hold key, speak, release, done.
No internet needed. No audio uploaded. No server to depend on. Your voice stays on your Mac, and the model runs on your CPU.
100% offline
Data sent to the cloud
From the moment you stop speaking to the moment text appears. No spinner, no delay. This is what real-time actually means.
No cloud means no latency from network calls, no privacy risks, and no dependence on servers that might be slow or down.
English, Chinese, Japanese, Korean, and Cantonese — all at the same 70ms speed, all processed locally.
Slack, VS Code, Notion, Mail, Chrome — clipboard paste means it works everywhere. No app-specific plugins.
Click any app below to see how real-time voice to text works in it.
One-time purchase. No subscription. No cloud.