Pulse Record
Install Pulse Record, grant microphone permissions, dictate into your workflow, and recover from transcription issues.
What Pulse Record is
Pulse Record is the macOS dictation companion for hands-free writing and workflow control. It records speech, transcribes it, applies dictionary and cleanup formatting, and helps turn spoken work into usable text.
Pulse Record is not Pulse Annotate and it is not Pulse Narrate. Record is the native dictation app. Narrate is the deliverable voiceover workflow.
Section outline and media plan
| Reader question | Pulse Record section | Media to use |
|---|---|---|
| When should I use it? | Dictation target workflow | Pulse Record app media while public dictation screenshots are finalized. |
| What does it produce? | Transcript review | Transcript and cleanup examples grounded in the formatter. |
| What permissions are needed? | macOS permissions prompt | Info.plist microphone and speech recognition strings. |
| How do I recover when it fails? | Permission, input target, and transcript recovery | Help Center recovery links for microphone and wrong-target issues. |
![]()
When to use it
Use Pulse Record when speaking is faster or less disruptive than typing. It is for dictated input: notes, status updates, rough drafts, issue descriptions, meeting capture, and hands-free text entry into the focused app.
Do not use Pulse Record to annotate a webpage or generate a deliverable soundtrack. Use Pulse Annotate for visual feedback loops. Use Pulse Narrate for audio proof generated from Pulse AI-written response or deliverable text.
Dictation target workflow
Pulse Record follows the active macOS input target. Before recording, click the exact text field, document, editor, or task body that should receive the transcript. Keep focus stable while dictation is active.
The safest workflow is:
- Put the cursor in the intended target.
- Record a short spoken chunk.
- Stop and review the transcript.
- Apply dictionary corrections when repeated terms are wrong.
- Move to the next target only after the current chunk is complete.
This avoids the most common wrong input target failure: focus moves after recording begins, so text lands in another app or field.
What it produces
Pulse Record produces text. The output is a transcript that can be inserted into the target app, reviewed by the user, corrected with dictionary entries, and cleaned up by the formatter for numbers, dates, times, currency, and ordinals.
Treat the transcript as drafted input, not final proof. Review it before sending it to an agent, customer, task, or launch artifact.
Install
Pulse Record ships as a macOS app named Pulse Record. The current app bundle metadata uses CFBundleDisplayName and CFBundleName set to Pulse Record, while retaining the existing internal executable identifier for update and signing continuity.
Use the signed release package supplied by the Pulse AI release process. During launch prep, do not treat the website "download" button as proof that a public installer is available; the public download path must point to a real release artifact before it is called shippable.
After install:
- Open Pulse Record.
- Keep it available while you work in the target app.
- Confirm macOS permission prompts when they appear.
- Run a short test dictation before relying on it for a long session.
Permissions to expect
macOS controls recording access. Pulse Record declares:
NSMicrophoneUsageDescription: Pulse Record uses your microphone to capture and transcribe speech.NSSpeechRecognitionUsageDescription: Pulse Record uses speech recognition to transcribe recorded audio.
If recording does not start, open macOS System Settings and check Microphone and Speech Recognition permissions for Pulse Record. After changing permissions, quit and reopen the app so macOS applies the new grants cleanly.
Those permission strings are source-backed in the native recorder Info.plist as NSMicrophoneUsageDescription and NSSpeechRecognitionUsageDescription.
Dictation workflow
Use Pulse Record for spoken drafting, status capture, notes, and recovery when typing would interrupt your flow.
- Start recording.
- Speak in complete chunks.
- Pause between sections when you want cleaner segmentation.
- Stop recording.
- Review the transcript before sending or saving it.
- Correct custom vocabulary through the dictionary when repeated phrases are wrong.
Shorter recordings are easier to review and recover. For important work, dictate in sections instead of one long uninterrupted pass.
Dictionary behavior
Pulse Record includes a custom transcription dictionary. The dictionary stores heard-to-corrected phrase pairs in user defaults. A fresh install can include defaults such as converting "bot to bot" into "Pulse Record".
Use the dictionary when the recognizer repeatedly mishears product names, proper nouns, project names, or domain phrases. Dictionary corrections apply after transcription, so they are best for stable replacements rather than one-off edits.
Dictionary entries can be represented as comma-separated pairs such as heard phrase→Correct Phrase or heard phrase=Correct Phrase.
Cleanup behavior
Pulse Record also includes formatter cleanup for spoken numbers, dates, times, currency, and ordinals. The formatter can convert spoken phrases such as dollar amounts, dates, and times into cleaner written text.
Cleanup is meant to improve readability without changing the intent of what you said. Always review output before using it in a task, customer message, or launch artifact.
Recover when it fails
If recording fails, check macOS permissions first. If transcription succeeds but text is poor, add dictionary corrections for repeated product terms and try a shorter recording. If cleanup changes text incorrectly, disable formatter behavior for that pass or manually restore the phrase before sending.
If the app appears open but does not respond, quit Pulse Record and reopen it from Applications. If macOS recently updated permissions, a restart can also clear stale permission state.
If a transcript is important, copy it into the target task or note immediately after review. Do not rely on a temporary recording session as the only copy of work.
If dictation lands in the wrong input target, stop recording, click the intended field again, and test with one short phrase before continuing.
Troubleshooting belongs in Help Center once the product behavior is clear. Start with Pulse Record support, Pulse Record cannot hear my microphone, and Pulse Record is typing into the wrong place.
Next: Pulse Narrate, API Keys & Providers, and Preferences.