To turn an iPhone Voice Memo into text, check whether Voice Memos already offers a transcript. Copy it if that solves the task. For Vox.md’s on-device transcription and Markdown routing, export the recording to Files, import that audio into Vox.md, review the transcript and save it to your chosen folder.
This guide starts with an existing recording in Apple Voice Memos. For a new recording made directly in Vox.md, use the voice notes to Obsidian guide. The walkthrough is based on documentation and source review, not a newly performed device export.
1. Check whether Voice Memos already has the text
Apple documents transcription on supported iPhones, languages and regions. In Voice Memos, select the recording, open its More menu and look for View Transcript or Copy Transcript. Copy the full transcript or selected text, then paste it into your note if that is all you need.
You do not need another app just to copy an available Apple transcript. Consult Apple's transcription guide for current availability. A missing option can reflect device, system, language or regional support; it does not establish that the audio is damaged.
2. Export audio when you want Vox.md to transcribe it
Apple's Voice Memos export instructions describe selecting a recording, opening More → Share, and choosing Save to Files. Choose a folder you can find again, then save.
The normal export is an .m4a recording. Options can preserve editable layers or effects in a different representation. Start with the ordinary flattened audio export for this workflow; this guide does not establish that every layered or editable export works with every transcription backend. Keep the original Voice Memo.
Sharing an audio file is different from copying Apple's transcript. Importing the audio into Vox.md runs its selected transcription backend again; it does not import Apple's transcript or guarantee the same wording.
3. Prepare a backend and a Markdown destination
- In Vox.md, choose a transcription backend that supports your device and recording language. Complete any required model download while connected.
- Create or select a Capture Preset with a writable Markdown folder. Use New Note for the first import, so the resulting file is easy to identify.
- Choose Keep Original processing when you want the recognized text without an optional AI rewrite.
- In Quick Capture's voice controls, select Add to Draft before importing, so you can review the first transcript.
Keep Original preserves the recognized text; it does not make recognition verbatim or error-free. Automatic, Apple Speech and downloaded models have different device and language requirements. Check the model reference and your installed app's current purchase or access requirements before a longer import.
For Obsidian, choose the actual vault folder exposed through Files. Vox.md does not connect directly to the Obsidian Sync service. A plain writable Markdown folder works too.
4. Import the saved recording and review the draft
- Open Import audio or video in Quick Capture's voice controls.
- Select the exported recording from Files and allow the import and transcription to finish.
- With Add to Draft selected, review the transcript in the composer. Correct names, dates, numbers and other important wording against the source audio.
- Confirm the selected preset and destination, then tap Send.
Use the transcription import control rather than attaching the file through a generic Files attachment action. An audio attachment in a note is not itself a transcript.
With Send Immediately selected instead, a completed transcript can proceed through the selected preset without the same draft-review step. Set Add to Draft first for this initial check. Finish any active recording before starting an import.
5. Check the actual Markdown file
Wait for delivery to complete. Open the destination in Files or Obsidian and confirm the file, text and filename. Seeing a transcript in the composer proves recognition produced text; it does not prove the folder write succeeded.
A synthetic recording might produce this note body:
Review the garden plan on Saturday.
Bring the updated measurements.
This is an illustrative output, not a real recording or accuracy test. Your preset controls the filename, metadata and formatting. A new import's date or filename is not automatically the original Voice Memo's recording date. Add that context deliberately when you need it.
After one new-note import works, follow the daily-note guide for a rolling destination or the template recipes for an inbox or task entry. Match the intended note's date and path before sending older recordings into a daily journal.
Limits, privacy and failed imports
- Import is unavailable: finish the current recording, check your app's access requirements, and confirm the selected preset is enabled.
- Recognition fails: check backend readiness, language, local storage and whether the file's audio can be decoded. Keep the original and try a small ordinary audio export before a long or unusual file.
- Delivery is pending: restore folder access and retry the retained failed delivery. Reimporting the same memo can create a second capture.
- You need the original audio too: configure and verify audio retention separately. Do not assume every preset saves an audio attachment merely because it produced text.
Vox.md runs transcription on device. Initial backend preparation can require a download, and a Files provider can need connectivity to retrieve the source or write the destination. If you save audio or Markdown in a synced folder, that provider handles its own synchronization. For a complete disconnected workflow, follow the offline preparation and verification guide.
This route is not automatic Voice Memos library synchronization, a batch archive importer, or a promise to preserve Apple edits, layers, transcript timing or recording metadata. Start with one recording, keep its original, and verify the saved text before extending the workflow.