Every voice memo, transcribed.
Every time.
Record on your Apple Watch, iPhone, or Mac. My Transcriber picks it up, transcribes it with Whisper AI, and puts the text in your folder as markdown — searchable, shareable, and ready for your AI tools. 99+ languages. Runs locally. Free.
Mac: macOS 14+, Apple Silicon (M1–M5, fast) • Also available for Intel (slower, but it works)
Windows: the Transcribe screen (record, or give it a file); Voice Memos are on the Mac.
Free download.
You already record voice memos.
Now make them useful.
Apple Voice Memos is great for recording. But the built-in transcription? Only ~12 languages, doesn't always work, and the text stays trapped inside the app with no way to export it. My Transcriber fixes all of that.
It just shows up in your folder
Record a voice memo. Walk away. The transcript appears in your folder as a markdown file, organized by date. Background processing catches up even if you recorded while offline or on the go. You don't have to think about it.
Record on your Apple Watch during a walk. By the time you're at your desk, it's transcribed.
Live transcript while recording
In a workshop or meeting? The transcript writes to the file as people speak. Open it alongside your notes to check what someone said five minutes ago — without stopping the recording.
"Wait, what was the deadline?" Just check the file.
Your team's transcripts, one folder
Point My Transcriber at a shared Dropbox or SharePoint folder. Each person gets their own account subfolder. Everyone's transcripts are organized, searchable, and accessible to the whole team. No special software needed to read them.
Markdown files. Anyone can open them.
99+ languages. Automatically.
Apple Intelligence handles a handful of languages. Whisper handles 99+ with automatic detection. Speak Swedish in a standup, get Swedish text. Switch to English mid-sentence, it follows. No settings to change.
Essential for anyone not working exclusively in English.
Fast enough to be invisible
One hour of audio transcribes in roughly two minutes on Apple Silicon — about how long it takes to make a coffee. Intel Macs are slower (closer to real-time), but they work too. Drop a file, glance away, the transcript is there.
Ways people use it
Any time you'd record a voice memo, My Transcriber makes sure the text ends up somewhere useful.
Meeting notes
Hit record on your Apple Watch. The transcript shows up in your team's shared folder with timestamps. No recording software to install.
Phone calls
Both people record on their own device. Both transcripts land in the shared folder with timestamps — reconcile them later into a complete record.
Quick thoughts
Record a voice memo while walking. By the time you sit down, the thought is already a text file in your folder, ready to act on.
Workshops & lectures
Live transcript builds up as people speak. Open the file during the session to reference what was said. No stop-and-copy cycle.
Podcast drafts
Each host records their own voice memo. Transcripts with timestamps make it easy to edit and compile without dedicated podcast software.
AI context
Drop transcriptions into Claude Projects or ChatGPT for summarization, task extraction,
or follow-up. Your voice becomes structured input for any AI workflow — and
scripts or agents can feed audio in directly with the
transcriber transcribe CLI, which imports the file into your
library with live progress in the app.
How It Works
Works with Apple Voice Memos. No configuration needed.
Record
Hit record in Apple Voice Memos — on your Mac, iPhone, or Apple Watch. Or use the in-app push-to-talk button.
It transcribes
My Transcriber detects the recording and transcribes it locally with Whisper AI. Live if you're at your Mac, or catching up in the background when you return.
Text in your folder
The transcript is a markdown file organized by date. Open it, share it, search it. Accept it to delete the audio, or keep both.
No cloud. No accounts. No API keys. Just download and go.
Using it with Claude Code
Claude's own voice input only listens while you're sitting in front of it, and it dictates twenty languages where Whisper handles 99+. A recording you already made isn't something you can hand Claude either. Once it's a markdown file in your folder, Claude Code can read it, search it, and act on it like anything else you've written.
Talk while you work
Hit record, switch to another app, review something, and talk out loud about it for minutes on end. Then paste the transcript back into Claude Code. The recording doesn't care which window is in front, or that you stopped to read a diff in the middle.
Record away from the desk
Record a voice memo on your Apple Watch, either your own thoughts or a conversation or a meeting, and as soon as it syncs over iCloud it gets transcribed on your Mac. By the time you sit down the transcript is already there.
Or skip the pasting
Install the My Transcriber plugin in Claude Code or in Codex, and you can say "I talked
about this in one of my latest voice memos, check it out" in any session and it goes and reads
them itself. My Transcriber needs no AI subscription of its own; the plugin uses the
one you have: Claude (Anthropic) or ChatGPT (OpenAI, through Codex). It also ships an MCP
server (transcriber mcp-server) for any other MCP-aware tool.
How It Compares
| Feature | My Transcriber | MacWhisper | Wispr Flow | Apple Voice Memos |
|---|---|---|---|---|
| Automatic Voice Memos transcription | Yes | No | No | Partial |
| Live transcript to file | Yes | No | No | No |
| Background catch-up processing | Yes | No | N/A | No |
| Markdown output | Native | Export only | No | No |
| Shared team folder | Yes | No | No | No |
| Languages | 99+ | 99+ | 100+ | ~12 |
| Runs locally | Yes | Yes | No (cloud) | Yes |
| Delete audio, keep text | Yes | No | N/A | No |
| Price | Free | $30-80 | $12-15/mo | Free |
Runs entirely on your Mac
Whisper AI runs locally with Metal GPU acceleration. No cloud, no accounts, no data leaves your device. A 30-minute recording transcribes in under a minute on Apple Silicon.
Skeptical? We publish the benchmark script and a public-domain audio clip so you can run it on your own machine and compare.
100% offline
After the one-time model download. No internet needed.
No accounts
No sign-up, no telemetry, no data collection.
Your files
Plain markdown on your filesystem. No vendor lock-in.
Frequently Asked Questions
Everything you need to know about getting started
Latest from the blog
See all posts →v0.6: A New Look, and My Transcriber in Claude Code and Codex
My Transcriber 0.6.0 moves onto the same paper design as our other apps, gets a plugin so any Claude Code or Codex session can use it, comes as an early alpha for Windows, and fixes a handful of things, including voice memos it had been quietly skipping.
v0.5: A Shorter Setup
Getting started with My Transcriber is two permissions and a download, and it now says so on the first screen instead of letting you find out as you go. There's also a new option worth knowing about if you're on a recent Mac: Apple's own on-device speech engine, roughly twice as fast on long recordings and with no model download at all.
v0.4: Instant, Self-Healing Transcription
Fresh recordings transcribe the moment they appear, the background helper heals itself instead of asking you to quit and reopen, and the CLI now imports recordings straight into your library. Plus a dramatically smaller download.
Voice memos are easy to record.
Now they're easy to use.
Download My Transcriber for Mac. Free, offline, 99+ languages.
Every voice memo becomes a markdown file you can open, share, search, and feed to AI.
Apple Silicon (M1–M5) • Also available for Intel