SapoWhisper
GitHub Download
View on GitHub Download for macOS
macOS 14+FreeLocal or cloudBring your own keys

Native dictation for macOS

Trigger recording from anywhere, transcribe with Apple, WhisperKit, Google Cloud Chirp 3, Deepgram, or ElevenLabs Scribe Realtime v2, then let Gemini polish the result before it pastes itself into whatever you're writing.

Download for macOS View on GitHub
Press ⌥ + Space to start dictating
Auto-paste · Searchable history · Gemini polish
Open source on GitHub · MIT licensed
SapoWhisper Ready to record
⌥ + Space
Last transcript

“Can you send the updated launch copy before lunch? Also, push the release branch to GitHub.”

Auto-paste text Pastes automatically when done
History
Settings ElevenLabs Realtime
Quit SapoWhisper
Transcription engines

Pick the right tradeoff per moment

Local, cloud, and realtime engines share one workflow. Stay local for privacy, swap to the cloud for accuracy, or stream in real time — without leaving the app.

Recommended

Whisper Local

WhisperKit on-device

100% private. Audio never leaves your Mac. Works without an internet connection.

  • Local
  • Free
  • Offline

Apple Speech

Online · Apple servers

Free and ready to use. Sends audio to Apple's recognition service when online.

  • Cloud
  • Free
  • Fast

Google Cloud Chirp 3

Bring your own GCP project

Maximum precision. Uses Google Cloud Speech-to-Text with your own credentials.

  • Cloud
  • Highest accuracy
  • GCP

Deepgram Nova-3 / Flux Live

Batch or real-time streaming

Nova-3 for high-accuracy batch transcription, or Flux Live for low-latency streaming with WAV backup.

  • Cloud
  • Streaming
  • Low latency

ElevenLabs Scribe v2 / Realtime v2

Scribe batch or WebSocket

Use Scribe v2 for accurate batch transcription, or Scribe Realtime v2 for committed-text streaming.

  • Cloud
  • Realtime
  • Scribe
AI improvement

Gemini polishes every transcript before you paste

SapoWhisper runs each transcript through Gemini 3.1 Flash-Lite on Vertex AI. Pick a mode, plug in your project context, and get a clean message instead of a raw dictation dump.

  • Prompts per project

    Create editable modes for Codex, Claude Code, Slack, issues, or any dictation flow.

  • Custom vocabulary

    Add product names, commands and frequent corrections so Deepgram and Gemini stop guessing.

  • Personal context

    One profile that explains your role and tools — applied to every AI mode without inventing details.

Powered by Vertex AI · gemini-3.1-flash-lite
Mode Work Message
Raw transcript

uh send to the team like the latest mockups before lunch i think and also push the release branch to github and we can talk after standup

Polished by Gemini

Sending the latest mockups to the team before lunch. Also pushing the release branch to GitHub. Let's catch up after standup.

Built-in modes
  • Automatic Balanced cleanup for everyday dictation
  • AI Assistant Prompt Turn dictation into a clear request
  • Work Message Polish Slack, email, and team threads
  • Translate to English Convert any transcript into English
Power features

Built around the real dictation workflow

Engines and AI handle the transcript. These are the everyday touches that make SapoWhisper feel like part of your Mac, not a separate tool.

Trigger anywhere

Global hotkey workflow

Press ⌥ + Space (or your own shortcut) from any app — the recording overlay appears without stealing focus.

Output

Auto-copy and auto-paste

The polished transcript lands in your clipboard and drops directly where your cursor is, so dictation feels like typing.

Audio control

Built for daily mic use

Preferred microphone sync, mic test, gain controls, audio cues and auto-ducking that lowers system volume while you talk.

Bilingual UI

Spanish and English

Swap the interface and the transcription language between Spanish, English or Auto — recognise both without restarting.

History

More than one-shot transcription

SapoWhisper keeps working after the paste. Search past transcripts, replay the original audio, pin important entries, and run the same clip through another engine when you want a better result.

Search what you already said

Find older transcripts quickly instead of dictating the same thing twice.

Replay the original audio

Listen back before sharing, editing, or comparing transcription quality.

Try another engine later

Re-process the same recording when accuracy matters more than speed.

Pinned transcript

Review, replay, and refine

Keep the original audio, compare engines, and export the result you actually want to keep.

Audio playback Re-transcribe Download audio Re-polish with AI
Recorded from preferred microphone
00:18 / 00:34 Deepgram Nova-3 · AI polished
Workflow

Five stages from voice to clean text

Trigger, capture, transcribe, refine, paste. SapoWhisper hides the complexity so you can stay in your editor.

  1. 01
    Trigger

    Press your shortcut

    ⌥ + Space (or your own) brings up the recording overlay from any app.

  2. 02
    Capture

    Speak naturally

    Preferred mic, gain control, auto-ducking and pause / resume when you need it.

  3. 03
    Transcribe

    Pick your engine

    Apple, WhisperKit, Google Cloud Chirp 3, Deepgram Nova-3 / Flux Live, or ElevenLabs Scribe Realtime v2.

  4. 04
    Refine

    AI polishes the text

    Gemini 3.1 Flash-Lite on Vertex AI cleans up the transcript without inventing details.

  5. 05
    Output

    Paste and keep history

    Auto-paste lands the result where your cursor is. Everything stays searchable.

Download

Ready to make dictation part of your everyday workflow?

Pick the transcription engine that fits the moment, let Gemini polish the result, and start writing with your voice from anywhere on macOS.

  • macOS 14.0+
  • Apple Silicon only (M1+)
  • Free and open source
  • Optional: your own GCP, Deepgram, or ElevenLabs keys
Download latest release macOS app View source on GitHub Report an issue