How to Transcribe a Voice Memo: Every Way That Works (2026)
iPhone has a built-in option. It doesn't work for everyone. Here's what to do when it does — and when it doesn't.
.m4a file off the phone and upload it to a transcription service. DeluxeScribe handles voice memo files in 99 languages with automatic language detection, 60 minutes free, no credit card. Below: every device path that works, the gotchas Apple Support doesn’t cover, and honest accuracy expectations.- 60 minutes free
- No credit card
- 99 languages
- Speaker labels
Last verified June 25, 2026
Voice memo vs voicemail vs M4A — quick disambiguation
Google conflates these three searches. They’re different problems.
- Voice memo — a recording you madeon your phone using the Voice Memos app (iPhone) or a Voice Recorder app (Android). That’s what this page covers.
- Voicemail — a message someone left you. Different app (Phone → Voicemail), different transcription path, different carrier dependencies. See our voicemail-to-text guide.
- M4A — the file format both end up in (AAC audio inside an MP4 container). If you already exported the file and want to know what M4A is and how to transcribe any M4A file regardless of source, see our M4A-to-text guide.
TL;DR — pick your path
| If you have… | Best path | Cost |
|---|---|---|
| iPhone 12+ on iOS 18, new recording, supported language | Apple Voice Memos built-in | Free, on-device |
| iPhone 11 or older — any recording | Export → upload to a service | Free tier → ~$10/mo |
| iPhone 12+ but recording predates iOS 18 | Export → upload to a service | Free tier → ~$10/mo |
| Apple-unsupported language (Turkish, Arabic, Hindi, etc.) | Export → multi-language service | Free tier → ~$10/mo |
| Pixel 3 or newer | Pixel Recorder app built-in | Free, on-device |
| Samsung or other Android | Voice Recorder app + export | Free tier → ~$10/mo |
| Recording > 30 minutes | Export → cloud service (built-in struggles on length) | Free tier → ~$10/mo |
| Sensitive content (medical, legal, confidential) | Self-hosted Whisper (no upload) | Free, requires Python |
iPhone — Apple Voice Memos transcription
Apple shipped on-device transcription for Voice Memos in iOS 18 (September 2024). It’s free, runs on the device with no network call, and produces a read-only transcript you can copy out.
Requirements
- iPhone 12 or newer(the Neural Engine on earlier models isn’t powerful enough for on-device transcription)
- iOS 18 or later installed
- Recording made after iOS 18 was installed — this is the gotcha (see below)
- Recording is in one of the 10 supported languages
Supported languages (iOS 18 launch)
English, French, German, Spanish, Mandarin Chinese, Cantonese, Japanese, Korean, Portuguese, Italian. Apple expands this list over time — check the current list in Apple’s support documentation for your iOS version.
How to view the transcript
- Open Voice Memos
- Tap the recording you want to read
- Swipe up on the title bar (or tap the small “Aa” icon if shown) to expand the transcript
- The transcript is read-only — long-press to copy the text
The gotchas — when iPhone’s built-in fails
This is the section Apple Support and competitor SaaS pages don’t cover. If any of these apply to you, the built-in option won’t work and you need the export path:
- iPhone 11 or earlier.No on-device option. No software update will add it. The Neural Engine on the A13 chip and earlier doesn’t support it.
- Recording predates your iOS 18 install. Apple processes new recordings as they’re saved. Pre-existing memos aren’t backfilled. There’s no official “transcribe now” button for old recordings.
- Language not on Apple’s list. Turkish, Arabic, Hindi, Dutch, Polish, Vietnamese, Thai, and most others are excluded. The recording transcribes to nothing (or to a related Latin-script approximation).
- Recording is too long. On-device transcription works on long files but can pause, fail to complete, or drain the battery while processing. For recordings over ~30 minutes, the export path is more reliable.
Android — Pixel Recorder and other voice apps
Pixel Recorder (Pixel 3 and newer)
Google ships the Recorder app on every Pixel since the Pixel 3. It transcribes on-device, free, in English plus a growing list of additional languages depending on Pixel generation. Newer Pixels (Pixel 8, 9, 10) support more languages and offer speaker labels for multi-speaker recordings — a feature iPhone’s Voice Memos doesn’t have.
Samsung Voice Recorder
Samsung’s built-in Voice Recorder has a “Speech-to-text” mode (sometimes called “Voice Memo” depending on the Samsung version) that transcribes live or after the fact. English coverage is solid; non-English support is more limited than Pixel.
Other Android phones
If you’re on a OnePlus, Xiaomi, Motorola, or other non-Pixel Android device, you don’t get a built-in transcription option. Two paths:
- Install a third-party recording app with transcription (Otter, Transcribe, AudioPen) and re-record or import.
- Use your phone’s standard voice recorder, then export the file and upload to a service.
Export a voice memo from your phone
iPhone — three export paths
- Share button (simplest). Open Voice Memos → tap the recording → tap the three-dot menu → Share. You can AirDrop to a Mac, save to Files, email, message, or send to any installed app. Export format is
.m4a(AAC audio in MP4 container). - iCloud Drive sync.Enable iCloud Drive → Voice Memos in Settings → [your name] → iCloud → toggle Voice Memos on. Recordings appear in iCloud Drive’s Voice Memos folder, visible from any Mac or via iCloud.com.
- Lightning / USB-C cable. Plug the iPhone into a Mac and open the Voice Memos app on macOS — recent versions sync automatically. On Windows, use the Apple Devices app or iTunes.
Android — export from Recorder or Voice Recorder
- Open the recording in your recorder app → tap Share or Save
- Choose a destination (Google Drive, email, AirDrop’s Android equivalent “Quick Share,” or any installed app)
- File format varies by app — usually
.m4aon Samsung,.mp3on some others, or.amron older devices
The .m4r vs .m4a gotcha
If you’ve used a memo as a ringtone, iOS may save it as .m4r instead of .m4a. They’re the same audio format with a different extension. If a transcription service refuses an .m4r file, rename it to .m4a and try again.
Upload and transcribe a voice memo file
Any service that accepts .m4a works (most do). Honest options ranked by what each is best at:
| Service | Free tier | Paid from | Best for voice memos when… |
|---|---|---|---|
| DeluxeScribe | 60 min one-time | $10/mo · 1,200 min | Multi-language memos, batch of old recordings, cheapest per-minute |
| Apple Voice Memos (built-in) | Free, on-device | Free | Short English memos on iPhone 12+ that you just recorded |
| Pixel Recorder (built-in) | Free, on-device | Free | Pixel users with multi-speaker recordings (has speaker labels) |
| Otter | 300 min/mo | $17/mo | Meeting-style memos with notes, calendar integration |
| Self-hosted Whisper | Free | Free | Sensitive content you don’t want to upload |
Step-by-step with DeluxeScribe
- Sign up (60 minutes free, no credit card)
- Drag your
.m4ainto the upload area, or click to browse - Language is auto-detected; speaker labels are optional (skip if it’s just you talking — they cost a small amount of accuracy)
- Wait 1-3 minutes for a typical voice memo (longer for files over an hour)
- Review the transcript in the browser editor, fix any mis-heard names or numbers, export as TXT, DOCX, PDF, SRT, VTT, or JSON
Accuracy reality for voice memos
Voice memos are usually easier for AI transcription than meeting audio or phone calls. Why?
- Close-mic. You held the phone near your mouth. Signal-to-noise ratio is high.
- Mono, single speaker.No diarization problem — the model doesn’t have to figure out who said what.
- Native phone audio sampling. iPhone records voice memos at 44.1 or 48 kHz uncompressed-ish AAC, not the 8 kHz narrow-band codec that destroys phone call quality.
Realistic accuracy expectations
| Recording condition | Typical word accuracy |
|---|---|
| Quiet room, phone close to mouth, clear English | 95-98% |
| On the go (walking, mild background noise) | 85-92% |
| In a car (engine, wind) | 70-85% |
| Multi-speaker meeting recorded as voice memo | 75-85% words; 60-80% speaker attribution |
| Heavily accented speech in supported language | 80-92% |
| Recording with music playing in background | 50-80% (model often confuses lyrics with speech) |
The same error patterns appear across every provider: phone numbers, proper nouns, technical jargon, and uncommon brand names are the most frequently mis-heard. Always spot-check these before relying on the transcript for a callback, a quote, or a citation.
Common gotchas
- Old recordings don’t get auto-transcribed on iPhone after iOS 18. This is the #1 complaint in Apple Discussions and TidBITS threads. No official workaround; export and upload is the path.
- Apple’s transcript is read-only inside Voice Memos.You can’t fix mis-heard words in the app. Copy the text into Notes or another editor for cleanup, or use a third-party service that gives you an in-browser editor.
- Battery drain on older devices. On iPhone 12 mini and iPhone 13 mini, on-device transcription of long files visibly drains battery. Plug in for files over 30 minutes.
- AirDrop file-size limits. Very long recordings (multi-hour) can fail to AirDrop. Use iCloud Drive sync or USB cable for those.
- iCloud Drive sync delay.Voice memos don’t always sync immediately. If a memo isn’t appearing on your Mac, force-quit Voice Memos on iPhone and reopen to trigger sync.
- Transcript may differ from displayed text. When you copy from the Voice Memos transcript view, you get the version Apple finalized after processing. The live-transcription view during recording can show different (less accurate) text.
How this page was verified
Related guides
- Voicemail to TextDifferent audio: voicemails are messages others left you, not recordings you made. Covers iPhone Visual Voicemail and the carrier paths.
- M4A to TextAbout the file format itself. Useful if you already exported and want to know what M4A is and how it compares to MP3.
- How to Transcribe AudioThe broader pillar — every path (SaaS, free tools, self-hosted Whisper, native OS) and how to pick.
- Podcast TranscriptionIf you recorded an interview as a voice memo for podcast production — speaker labels, show notes, and the publishing spec.