wispa

Private by design

Your words never have to leave your computer.

Choose a local model and wispa transcribes everything on-device with Whisper. No audio upload, no account, no telemetry. Want cloud speed or AI clean-up? Bring your own key.

  • On-device by default
  • Works offline
  • No account, no telemetry

How it stays private

Local by default, cloud only if you ask.

  1. 01

    Pick a local model

    Choose Whisper, plus Parakeet on Apple Silicon, and wispa downloads it once so transcription runs on your own hardware.

  2. 02

    Your audio is processed on-device

    Speech is transcribed locally and the audio is discarded on your computer. Nothing is uploaded, and there is no account to sign into.

  3. 03

    Cloud stays your decision

    If you ever want cloud speed or AI clean-up, you opt in with your own OpenAI, Groq or Anthropic key. It is off until you choose it.

What that means for you

Privacy that is the default, not a setting.

  • Nothing leaves your machine

    With a local model, your voice and the text it becomes stay on your computer. There is nothing to log into and nothing uploaded.

  • Works fully offline

    Local models run without a connection, so you can dictate on a plane, offline, or on a locked-down network with no loss of function.

  • No account, no telemetry

    wispa needs no sign-up and collects no usage data. You are not the product, and there is no profile building up in the background.

  • Cloud is opt-in, with your key

    Prefer a cloud model or AI clean-up? Add your own API key. You stay in control of what leaves your device and what does not.

FAQ

Privacy, answered.

Your words, your computer

Dictate without sending your voice away.

Download wispa, pick a local model, and transcribe everything on-device. No account, no telemetry, cloud only if you ask. Free while in beta on Mac and Windows.