Voice memos are great: they’re quick to capture and don’t require my eyes being glued to a screen. However, as audio files they don’t integrate well into my Obsidian notes.

I thought: Wouldn’t it be cool to take voice memos, and have a perfect transcript + summary automatically show up in Obsidian?

It turned out to be easier than I expected. No cronjobs or even API keys required.

• Apple Voice Memos
• MacWhisper
• Apple Shortcuts
• ChatGPT

Basically, it works like this:

The Logic:

1. Record in Voice Memos (iPhone or Mac).
2. MacWhisper watches the iCloud sync folder and auto-transcribes the audio.
3. An automation triggers a Shortcut.
4. ChatGPT cleans the text and generates a summary.
5. The final note is templated and saved into Obsidian.

Here is the high-level flow:

Setup §

1. Voice Memos §

Self-explanatory: record a new voice memo. Crucially, ensure iCloud sync is enabled so the files propagate to your Mac.

2. MacWhisper §

MacWhisper wraps Whisper and other models in a native UI. The killer feature here is the watch folder capability.

You need to configure it to watch the specific directory where the Voice Memos app stores recordings: ~/Library/Group Containers/group.com.apple.VoiceMemos.shared/Recordings

Whenever a new memo syncs to this folder, MacWhisper automatically generates a transcription .md file in the same directory.

3. Shortcut Automation §

With macOS Sequoia (and later), Apple added Automations to the Shortcuts app on Mac. This allows us to trigger actions based on folder events.

I set up an automation to run whenever a new file (the transcript) appears in that same Group Containers recording folder.

4. The Shortcuts §

I split the logic into three modular shortcuts for easier debugging, though you could combine them into one.

4.1 Move transcripts to Obsidian §

This simply moves the raw .md transcript files from the system folder into a transcripts/inbox folder inside my Obsidian vault to prevent processing duplicates.

-> iCloud link

4.2 Process Obsidian Inbox §

This is the heart. It:

1. Moves the transcript from inbox to transcripts/processed.
2. Passes the text to the Ask ChatGPT action.
3. Prompts ChatGPT to 1) Clean up the grammar/structure 2) Summarize the content 3) Suggest a filename
4. Splits the output.
5. Creates a new note in Obsidian (→ see example)

-> iCloud link

4.3 Master Trigger §

This combines the previous two. It checks if new files exist before running to keep things efficient. This is the specific shortcut targeted by the Automation in step 3.

-> iCloud link

Final Thoughts §

I capture most of my thoughts via voice now. It flows much better than typing on a phone.

I’m aware this setup relies on a few moving parts and could be solved with a custom Python script or various other tools, but this “no-code” approach feels remarkably stable. If you have ideas on how to optimize this, let me know!