Opus is the most common voice-message format in 2026. WhatsApp voice notes are Opus inside an OGG container. Telegram voice messages are Opus. Discord voice messages are Opus. If someone has sent you a 90-second voice message in the last week, you've been sent an Opus file. MDisBetter transcribes them directly so you can read instead of listening.
Why Opus is everywhere for voice messages
Opus was designed for two things: efficient voice compression at low bitrates (down to 6 kbps for usable speech) and low encoding/decoding latency for real-time use. WhatsApp, Telegram, and Discord all picked it because it does both better than any alternative — a 1-minute voice note is typically 30–80 KB at WhatsApp's default bitrate, vs ~500 KB as MP3 or ~5 MB as WAV. That size difference is what makes voice messages work over patchy mobile data; the same property is why Opus voice messages have largely replaced MP3 in messaging apps.
Practical workflows for voice-message transcription
Read instead of listen: the most obvious case. Save the WhatsApp voice note (long-press → Save), drop into MDisBetter, read the transcript in 30 seconds vs listening to a 2-minute monologue. Search across voice messages: if you save voice notes from a key contact (sales calls, client briefings, reporter source recordings), batching them through transcription gives you a searchable text archive of conversations that would otherwise be opaque audio. Quote extraction: for journalists or researchers, voice-message-as-source becomes citable text.
Note on Opus vs OGG
The file extension can vary — Opus audio commonly ships as .opus (bare Opus stream), .ogg (Opus inside an OGG container, common for WhatsApp/Telegram exports), or even .oga. MDisBetter accepts all three; the transcription pipeline detects the actual codec from the file header rather than the extension. See also OGG to Markdown for the broader OGG container.
Before / After
Before (PDF):
[Opus audio file]
Opus codec in OGG container, 38 KB, 47 seconds mono
Metadata: WhatsApp voice message export
(no extractable text — Opus-compressed waveform)
After (Markdown):
# Voice Message — May 9, 2026
Hey, just wanted to give you a quick update on the situation we talked about. So I spoke to the contractor and he confirmed they can start next Monday but they need the deposit by Friday at the latest. He also mentioned that the price went up a bit because of the materials, so it's now $14,500 instead of $13,800. Let me know if that's OK and I'll send him the deposit. Otherwise we can probably push it another week. Talk soon.
Frequently asked questions
Will WhatsApp voice notes transcribe correctly?
Yes — WhatsApp voice notes are Opus inside an OGG container, which MDisBetter handles directly. Save the voice note from WhatsApp (long-press → Save, or via WhatsApp Web → download), drop into the converter, get Markdown back. Single-speaker voice messages convert cleanly; very short clips (under 5 seconds) sometimes have less context for accurate transcription of unusual words.
How do I export a voice message from WhatsApp?
On mobile: long-press the voice message → Forward → Email/Share to yourself or to a cloud drive, the file saves as .opus or .ogg. On WhatsApp Web: hover the message, click the dropdown arrow, choose Download. The file ends up as Opus regardless of platform; MDisBetter accepts it directly.
Can I transcribe Telegram voice messages?
Yes — Telegram voice messages are also Opus (inside an OGG container, .ogg extension). Forward the message to "Saved Messages" → tap the file → Save to device → upload to MDisBetter. The workflow is the same as for WhatsApp.
Are Opus files at very low bitrates (6–12 kbps) still transcribable?
Yes, but accuracy degrades on the very lowest bitrates. WhatsApp's default voice-message bitrate is around 32 kbps, well within the comfortable range for accurate transcription. Below ~16 kbps, expect more confused words on technical terms or proper nouns; the overall transcript usually remains readable and the gist always survives.
Does this support push-to-talk recordings from gaming/voice apps?
Yes — Discord voice messages and most modern PTT apps record as Opus. The same transcription path handles them. For continuous voice chat (whole sessions of voice chat exported as one file), the transcript is one continuous block of text with sections and timestamps: we do not identify who is speaking, so a busy multi-person session is harder to follow than a single-voice clip.