WHISPERILA USER MANUAL · STEP-BY-STEP

The complete setup guide.
From installation to effortless speaking.

Learn how to install, configure, and master Whisperila on macOS. Get 100% private, on-device voice dictation running across every app on your Mac in less than two minutes.

01 Installation 02 Permissions 03 Global Hotkey 04 Local Models 05 Meeting Notetaker 06 Vocabulary 07 Verification
01 Step 1 · Installation

Download and Install Whisperila

Whisperila is distributed as a signed, Apple-notarized universal macOS disk image (.dmg). It contains zero telemetry, requires zero user account creation, and never asks for a credit card.

Apple Silicon (M1, M2, M3, M4)

Recommended for all modern Macs. Employs hardware CoreML Neural Engine acceleration for sub-120ms offline recognition.

Download Apple Silicon (.dmg)

Intel Mac (x86_64)

Built for older Intel-based Macs running macOS 14 Sonoma or newer. Runs local Whisper architectures via multi-threaded CPU execution.

Download Intel (.dmg)
Installation Step: Open the downloaded .dmg file and drag the Whisperila icon directly into your Applications folder.

First Launch on macOS: "Apple could not verify Whisperila"

Because Whisperila is an open-source, community-built application, macOS Gatekeeper may show a verification dialog on your first launch. Choose either quick method below to open it:

Easiest Method · No Terminal
  1. In Finder, open /Applications.
  2. Right-click (or Control-click) on Whisperila.app and choose Open.
  3. A prompt will appear with an Open button. Click Open. macOS will remember this permanently and never ask again.
Alternative · System Settings
  1. Go to System Settings → Privacy & Security.
  2. Scroll down to the Security section.
  3. Click Open Anyway next to "Whisperila was blocked from use".
Developer Shortcut · Terminal Command macOS 14+
Open Terminal and run this one-line command to remove the Gatekeeper quarantine attribute immediately:
Terminal — zsh
$ xattr -cr /Applications/Whisperila.app
02 Step 2 · Security Permissions

Grant Essential macOS Permissions

To provide seamless voice-to-text dictation that types directly wherever your cursor is active, macOS requires two straightforward privacy permissions:

Permission Why Whisperila Needs It Required?
Microphone Access Allows Whisperila to record your spoken words and process them locally through the on-device neural model. Mandatory
Accessibility Access Allows Whisperila to detect your currently active text field in any Mac application and paste your completed transcript automatically without manual keyboard copying. Mandatory
Screen & System Audio Recording Only needed if you use the AI Meeting Notetaker to capture system speaker audio from Zoom, Google Meet, or Teams alongside your own microphone. Optional
Navigate to System Settings → Privacy & Security to review or adjust permissions at any time. Whisperila never transmits audio or text data over the network.
03 Step 3 · Global Shortcut

Master the Global Hotkey

Whisperila operates as a discreet menu bar companion that you trigger instantly from anywhere across macOS with a simple keystroke:

Default Trigger Hotkey

⌥ Option + Space

Press once to start dictating. Speak naturally at your normal conversational cadence. Press again when you finish speaking to automatically paste.

Meeting Notetaker Trigger

⌥ Option + ⇧ Shift + Space

Starts continuous dual-channel recording for meetings and calls. Captures both mic and system audio until stopped.

Tip: In Settings → Shortcuts, you can customize the global trigger to any key combination or assign it to a dedicated function key (such as F5 or Fn).
04 Step 4 · Model Selection

Choose Your Offline Speech Model

Whisperila gives you full control over your local speech recognition engine. All models run 100% on-device with zero internet connection required:

Parakeet TDT CoreML

Engineered for maximum velocity. Delivers near-instantaneous ~110ms offline latency on Apple Silicon M-series chips with very low memory footprint.

Latency: ~110ms · RAM: 280MB · Offline

Whisper Large v3 Turbo

Ideal for multi-lingual dictation across 100+ languages and dense technical, legal, or biomedical terminology with maximum punctuation accuracy.

Languages: 100+ · Precision: Ultra · Offline
05 Step 5 · Dual-Channel Notetaker

Using the AI Meeting Notetaker

Unlike other applications that cap free meetings at 15 minutes or charge $15/month subscriptions, Whisperila includes unlimited meeting transcription for free:

1
Join Your Call: Connect to Zoom, Google Meet, Teams, or an in-person meeting.
2
Start Recording: Press ⌥ ⇧ Space or click "Start Meeting Notetaker" in the menu bar. Whisperila records your mic and the computer's output audio simultaneously on separate channels.
3
Auto-Generate Action Items: Click "Stop & Summarize". Whisperila generates structured speaker transcripts ("Me" vs "Them"), extracts key decisions, and compiles checklists ready to export to Notion, Slack, or Markdown.
06 Step 6 · Custom Vocabulary

Teach Whisperila Custom Vocabulary

Make dictation flawless for your specific workflow. Add specialized company acronyms, engineering terms, client names, and phonetic pronunciations in Settings → Vocabulary:

Examples of Vocabulary Rules

• "kubctl" → kubectl
• "whisper ila" → Whisperila
• "core m l" → CoreML
• "post gres" → PostgreSQL
07 Step 7 · Verification & Troubleshooting

Verify Your Offline Setup

Want to confirm that Whisperila works completely offline with zero cloud reliance? Follow this 30-second verification check:

  1. Turn off Wi-Fi on your Mac.
  2. Open TextEdit, Notes, or your code editor.
  3. Press ⌥ Space and speak a test sentence.
  4. Press ⌥ Space again. Notice your speech appears instantly with full punctuation and zero network activity.
Common Fix: If text does not appear in your active field, verify that Accessibility permission is toggled on under System Settings → Privacy & Security → Accessibility.