Dictation

Talk, and the words land where your cursor is

Tap ⌘⇧D to start and again to stop, or hold it to talk. Eavesy turns your words into text on your Mac and pastes it into the app you’re using. When you’re signed in, the audio also gets a cloud transcript, and that copy is lightly cleaned up with filler removed and punctuation added.

On-device first

The text pasted on your Mac is made on your Mac, with or without an account.

Cleanup in the cloud copy

Filler words and false starts are removed only in the cloud copy, never in the text pasted on your Mac.

Tap or hold

Tap to toggle, or hold for push-to-talk. Set Talk to fn if you like.

Dictate a technical instruction

Place the cursor in the target text field. Press ⌘⇧D to start Talk, speak the instruction, and press it again to stop. Holding the shortcut also supports push-to-talk. Pasting requires Accessibility permission; on-device recognition requires Speech Recognition access and available recognition support.

Say a short instruction first: “In the export handler, keep the workspace identifier separate from the selected row count.” Inspect the pasted text before sending it. Recognition can alter a product name, path, acronym, or identifier, so precision still needs review.

Local text and the cloud copy are different

The text pasted on your Mac normally comes from on-device recognition and can work without signing in. If on-device text is unavailable and you are signed in, the app can fall back to cloud transcription. A signed-in cloud copy may receive cleanup; that does not mean the already-pasted local text is silently rewritten later.

Do not assume a summary is produced for every short dictation. Talk is useful for entering text; completed meeting and video recordings have their own transcript and brief workflow.

Names and custom vocabulary

Signed-in transcription settings include custom words for names and jargon. Add the terms you actually use, then test them in a sentence. These hints help recognition but do not guarantee exact spelling or code syntax. Verify case, punctuation, and filenames before passing a dictated instruction to a coding agent.

If the microphone is audible but no words appear, check Speech Recognition availability separately from Accessibility. If no sound is recorded, start with the microphone troubleshooting guide.

Check your first technical prompt

  1. Put focus in the exact input where the words should land.
  2. Dictate one sentence with a project name or identifier you recognize.
  3. Stop and inspect the target field before submitting.
  4. Correct technical spelling and punctuation deliberately.
  5. If text is absent, distinguish an unavailable transcript from an unsuccessful paste.

If text exists in Eavesy but not in the target app, copying manually helps isolate the paste path. If saved audio is audible but no text is available, check recognition and processing. Neither result calls for repeating a long prompt before the short test works.

Capture it once. Use the context anywhere.

Capture on your Mac and keep normal files. Sign in for cloud transcripts and briefs, then share the context your AI needs.