WhispererWhisperer

Offline dictation for Mac. Code Mode for developers.

Hold Fn, talk, let go. Your words land in whatever app you're using. Nothing leaves your Mac.

Offline on-device transcription
Works in Slack, Gmail, VS Code, Terminal
Live preview while you speak

$14.99 lifetime — 17x cheaper than Superwhisper, no subscription like Wispr Flow.

Whisperer
Ready
About
Model
WhispererV3
547 MB
Microphone
MacBook Pro Microphone
Shortcut
Fn
Hold to record
How to use
1
Press Fn
2
Speak
3
Release

Why Whisperer

Three local engines (Whisper, NVIDIA, Apple Speech). Code Mode for camelCase and symbols. $2.99 base, $14.99 Pro lifetime.

Hold, talk, release.

Text goes straight into whatever field you're in. No clipboard dance.

Won't trigger by accident.

Fn+brightness or Fn+volume? Ignored. Only real dictation requests start recording.

Your voice stays on your Mac.

No accounts. No servers. No audio uploads.

Pro Pack Feature

For people who type a lot

Features for developers and anyone who dictates throughout the day.

Code Mode

Pro

Say symbols, casing, and structure the way you mean them in IDEs and terminals.

  • Say punctuation and symbols: parentheses, brackets, braces, quotes, commas, semicolons
  • Casing commands: camelCase, snake_case, CONSTANT_CASE
  • Literal mode turns off autocorrect for identifiers
  • Better spacing around operators and punctuation
Learn more about Code Mode
validator.py
1import re
2from typing import Dict
3
4def validate_input(data: Dict) -> bool:
5 """Validate user input data."""
6 if not data.get("email"):
7 return False
8
9 pattern = r"^[a-zA-Z0-9_.+-]+@[a-zA-Z0-9-]+\.[a-zA-Z0-9-.]+$"
10 return bool(re.match(pattern, data["email"]))
11
12def process_user(user_id: int) -> None:
13 """Process user data."""
14 user = get_user(user_id)
Code ModePython
UTF-8Ln 4, Col 42
Dictating...
Code Mode active
You said:
"def validate input open paren data colon dict close paren arrow bool colon"
Output:
def validate_input(data: Dict) -> bool:

Per-App Profiles.

Dictation that knows where you are.

Slack gets chat style. Gmail gets email format.
IDEs get code mode. It just switches.

Slack iconChat style
Gmail iconEmail style
VS Code iconCode style
Slack iconSlackChat Style

hey can you review the PR when you get a chance

Gmail iconGmailEmail Style

Hi team,
Please review the attached document.

VS Code iconVS CodeCode Style

def validate_input(data: dict) -> bool:

Switches automatically based on active app

How Whisperer works

Here's how it works.

01

Hold your key

Hold Fn or your chosen shortcut to start recording.

02

Speak naturally

See a live preview and waveform while you talk.

03

Release to insert

Let go. Whisperer cleans up the text and drops it where your cursor is.

Streaming during recording + a final pass on release for maximum accuracy.

Works in the apps you already use

Insertion uses Accessibility APIs with a paste fallback for compatibility.

Slack iconSlack
Gmail iconGmail
Google Docs iconGoogle Docs
Notion iconNotion
VS Code iconVS Code
Cursor iconCursor
Antigravity iconAntigravity
JetBrains iconJetBrains
Warp iconWarp
Notes iconNotes
Chrome iconChrome
Safari iconSafari
Slack iconSlack
Gmail iconGmail
Google Docs iconGoogle Docs
Notion iconNotion
VS Code iconVS Code
Cursor iconCursor
Antigravity iconAntigravity
JetBrains iconJetBrains
Warp iconWarp
Notes iconNotes
Chrome iconChrome
Safari iconSafari

What's inside

Everything runs locally on your Mac. No servers, no per-minute charges.

Fn
⌥

Hold-to-Record

Hold Fn (or your shortcut). Let go to transcribe and insert. Toggle and hands-free modes also work.

Live Preview
Thequickbrownfoxjumps

Live Preview

Words show up as you speak. ~300ms latency with dual-engine streaming.

Dictated

so basically we should go ahead and implement the caching layer

Rewritten

We should implement the caching layer to improve performance.

AI Post-Processing

Rewrite, translate, format, summarize, fix grammar. 10 built-in modes plus custom prompts. Local LLM, no cloud.

Whisper
Metal GPU
NVIDIA
Neural Engine
Apple Speech
System ML

Three Engines

Whisper (Metal GPU), NVIDIA (Neural Engine), Apple Speech. 10+ model sizes. Switch without restarting.

EN
ES
FR
DE
JA
KO
ZH
AR
PT
IT
RU
HI
+88 more

100+ Languages

Dictate in 100+ languages. Set different languages per app. Translate between them offline.

tenserflow→TensorFlowFuzzy
k8s→KubernetesExact
there→theirPhonetic

Personal Dictionary

Add your own terms. Three-tier correction: fuzzy, phonetic, and prompt words that boost recognition.

.mp3
Drop to transcribe

File Transcription

Drop an audio or video file to transcribe it. Same engine, local processing. No upload fees.

Weekly Activity
WordsTime
M
T
W
T
F
S
S
127
WPM
4.2k
Words
38
Sessions

History & Statistics

All transcriptions saved. Search, pin, flag, replay audio. See your WPM, peak hours, and which apps you use most.

um, so basically...
so basically...Filler removed
Basically...Grammar fixed
Basically...Ready

Text Processing

Auto filler word removal, spoken list detection, spell correction, trailing space handling.

Pro Pack (Lifetime)

For people who dictate all day. Code Mode, profiles, dictionary, and pro insertion.

Base App
$2.99 one-time

Core dictation for everyday use.

  • Hold-to-record dictation
  • Live preview
  • 100+ languages
  • Multiple Whisper models
  • Offline processing
  • File transcription
  • Filler word removal
MOST POPULAR
Pro Pack
$14.99 lifetime

Everything in Base, plus the good stuff.

Superwhisper: $249 · Voibe: $99 · Wispr Flow: $10–15/mo

  • Everything in Base
  • Code Mode
  • Per-app Profiles
  • Personal Dictionary
  • AI Post-Processing (10 modes)
  • Pro Insertion Engine

Code Mode

  • Symbol grammar
  • Casing commands
  • Literal identifiers

Per-app Profiles

  • Automatic switching
  • Different format rules per app
  • Per-app language (optional)

Personal Dictionary

  • Add names/acronyms/terms
  • Quick add from overlay
  • Import word lists (optional)

AI Post-Processing

  • Rewrite, translate, summarize offline
  • 10 built-in modes + custom
  • On-device LLM, no cloud

Pro Insertion Engine

  • Clipboard-safe paste
  • Typing fallback when paste fails
  • Per-app insertion method

No subscriptions. Pay once. Works offline.

One-time setup (required by macOS)

Whisperer needs macOS permissions to listen for your shortcut and insert text into other apps. This is standard for system-wide productivity tools.

Microphone

Used to record your voice.

Accessibility

Used to insert text into the active field.

Input Monitoring

Used to detect your shortcut globally.

Step 1
Install from App Store
Step 2
Grant permissions
Step 3
Ready to go!

Whisperer does not transmit audio. Permissions only enable local dictation and insertion.

Privacy by design

Transcription happens on your Mac
No account needed. No telemetry for core features
Models live in your Application Support folder
Diagnostic logs stay local (optional)

Contact support? You choose whether to attach logs. Nothing gets sent unless you send it.

Fast enough for daily dictation

Model preloading cuts first-dictation lag
Metal acceleration on Apple Silicon
Streaming with chunk overlap and deduplication
5 minute cap per recording keeps things stable

Frequently asked questions

Common questions about Whisperer.

Explore Whisperer

Speech to Text for Mac

Three offline engines. 100+ languages. $14.99 one-time.

Voice Coding

How to write code by voice on Mac.

For Developers

Code Mode, IDE support, RSI prevention.

Offline Dictation

Everything runs on your Mac. Nothing leaves the device.

Why Whisperer

How it compares to ChatGPT Voice, Superwhisper, and others.

Try it.

Pay once. Keep it forever. Nothing goes to the cloud.

Free trial included. Pro Pack $14.99 lifetime.