Anyone using voice memos → AI → Obsidian workflow?

Help
1

What I’m trying to do

I’ve been using voice memos as my primary idea capture method for the past three months, then feeding them into an AI pipeline before bringing them into my knowledge base.

My pipeline starts on iPhone and ends with Claude Code on Windows:

Voice memo → transcription → AI summary → structured notes → Obsidian

This works well for me for:

  • journaling

  • investment ideas

  • side projects

  • thinking-out-loud capture

However, I keep hitting one annoying issue: interruptions split recordings.

Examples:

  • phone calls

  • getting in/out of the car (Bluetooth switching)

Instead of one clean recording, I end up with more than one voice memos.

This breaks the pipeline because I now need to manually combine them before sending to AI.

Things I’ve tried

  • manually merging on desktop using audio editors (e.g., Audacity) — works but requires exporting files, aligning clips, and re-exporting

  • using video/audio editors like CapCut — import multiple recordings, place on timeline, export as one file (works but heavy for quick capture)

    These approaches technically solve the problem, but they add friction.

I’m curious

  • Are you using voice memos as input for Obsidian?

  • Do interruptions splitting recordings happen to you?

  • How are you handling it today?

  • Any recommended approach?

Would love to learn how others are doing voice-first capture.

2

Hey — I have almost the exact same pipeline. Voice memos on iPhone → transcription → AI summary → Obsidian folder. I use it daily for thinking-out-loud capture and project brainstorming.

The interrupted recording problem drove me crazy. Phone calls were the worst, but I also got splits from Bluetooth switching in the car. I’d end up with 2–3 files that were supposed to be one thought, and my transcription pipeline would process them as separate ideas. The context got lost.

I tried GarageBand too — they work but the friction killed the flow. Exporting to Files, importing to GarageBand, arranging on a timeline, exporting again… 5+ minutes for something that should take seconds.

I ended up building a small iOS app that solves just this one problem: Combine Voice Memos +. You select the recordings in Voice Memos, tap Share, and merge them right there. Takes about 10 seconds. The merged file goes straight back into my pipeline as a single clean recording.