Free & Open Source

Your voice.
Transformed.

Dictate into any app. Rewrite anything with AI. All private, all on your Mac. Nothing leaves your device.

Download — Free View on GitHub

v1.0.0  ·  macOS 26+  ·  Apple Silicon & Intel

Auto-Dictation Active  ·  Fn + Space to toggle

You've selected: "The meeting is scheduled for tomorrow afternoon"

Rewrite
Formal
Concise
Friendly

The floating toolbar · follows you across every app

Capabilities

Everything it does.
Nothing it doesn't.

Six capabilities. One lightweight app. No cloud required.

🎙️
On-Device Dictation

Apple's Speech framework transcribes your voice with precision. No internet, no servers, no data leaving your Mac.

AI Text Rewriting

Select text in any app and transform it using Apple's FoundationModels. Formal, concise, friendly — or your own instruction.

🌐
Works Everywhere

Via macOS Accessibility APIs, Mesmer works in every app — Slack, Xcode, Notes, Safari, everything.

Auto-Dictation Mode

Persistent listening without holding a key. Toggle with Fn + Space. A minimal pill UI confirms you're live.

📜
History

Every transcription and rewrite is saved locally. Review and reuse past sessions anytime — privately.

🔒
Complete Privacy

All processing stays on your Mac. No accounts. No telemetry. No cloud. Just your words, staying yours.

Two ways to use Mesmer.

01
Dictate into any text field

Place your cursor anywhere. Open Mesmer's floating toolbar, tap the mic, and speak. Your words appear exactly where your cursor is — instantly, without copy-paste.

02
Select → choose a persona → done

Highlight any text in any app. Mesmer's toolbar appears. Choose Formal, Concise, Friendly, or enter a Custom instruction. The selected text is replaced with the rewritten version — on-device, in seconds.

03
Review your history

Every session is logged locally. Open History to revisit, copy, or re-inject any past transcription or rewrite.

Rewriting Personas

One sentence.
Four voices.

Choose how your words sound. Mesmer transforms the meaning stays, the tone changes.

Formal
Professional & precise

"Hey, can we push the meeting?" → "I would like to request a reschedule for our upcoming meeting at your earliest convenience."

Concise
Short & direct

"I was thinking we could possibly maybe consider looking into this." → "Let's explore this."

Friendly
Warm & conversational

"Action items remain outstanding." → "Hey, just a heads up — we still have a few things to wrap up!"

Custom
Your instruction, exactly

Type anything. "Translate to Spanish." "Turn this into bullet points." "Make it sound like Hemingway." "Add more technical depth." Mesmer passes your instruction directly to the on-device model.

Requirements

What you need.

🍎
macOS 26.0 or newer
Required for Apple's FoundationModels framework. Dictation works on earlier releases, but AI rewriting requires macOS 26.
🎤
Microphone access
Granted once on first launch. Audio is processed entirely on-device — never transmitted anywhere.
Accessibility permissions
Needed to read selected text from other apps and inject transcribed or rewritten content back into them.
🗣️
Speech Recognition
A one-time permission grant. Used by Apple's native Speech framework for local, private transcription.

Build it yourself.

MIT licensed. Run the script, get a DMG.

build_dmg.sh
# Clone
git clone https://github.com/samirpatil2000/mesmer.git
cd mesmer

# Build — outputs Mesmer_Release.dmg
./build_dmg.sh
If macOS flags the app on first launch, run: xattr -cr /Applications/Mesmer.app

Speak your mind.
Beautifully.

Free. Open-source. Yours.

macOS 26.0+ 100% On-Device MIT License Free Forever