Dictation for Windows
Dictatray sits in the tray and turns your microphone into a keyboard. Focus a text field anywhere — a chat client, a terminal, an IDE, a form, a remote session — and talk. The words arrive as keystrokes, exactly as if you had typed them.
The product is the path from microphone to focused window. The engine that decodes your speech is your choice — two run on your machine, seven run on a provider you already pay. No server of ours ever sits in the audio path.
Dictatray lives in the tray and listens from startup. Mute is what stops it.
Speech detection cuts the pauses, not the words, and hands each phrase to the engine you chose.
The transcript is delivered character by character to whichever window has keyboard focus. Nothing to paste, nothing to switch to.
Free, offline and private. They cost nothing per use and work with the network unplugged — including a Portuguese fine-tune, which is not where the mainstream tools put their effort.
On your own API key, billed by your own provider. Choose one when you want its accuracy or its latency; the key is stored on your machine and used nowhere else.
Each engine carries its own language. pt is the default — auto-detect is not
one feature, and pretending it is costs you a sentence that changes language halfway.
By default, nothing. There is no account, no subscription and no telemetry. Pick a cloud engine and that engine's provider receives each utterance, on your key — a choice you make once and can see in the tray.
One file. Every library and every asset is linked inside it — there is nothing to install alongside it and nothing to uninstall.
dictatray.exeUnsigned build — Windows will warn on first run. Windows 10 (19045+) or 11, 64-bit.
The build is not code-signed yet, so SmartScreen shows "Windows protected your PC" on a machine that did not build it. That warning is honest — it means nobody has vouched for the file — and it goes away when signing ships.