Online Dictaphone

A recorder and transcriber for your voice memos, styled like a retro cassette recorder. Record as many takes as you like, watch your words appear live, and download whichever ones you want to keep.

THIS PAGE UPLOADS NOTHING · LIVE TRANSCRIPT USES YOUR BROWSER'S SPEECH SERVICE

Last updated: · Reviewed for accuracy against current browser APIs

Choose a language, then click REC and allow microphone access to begin.

Your words will appear here as you speak…

Diagram of one microphone signal splitting into a saved audio take and a speech-recognition transcript, which feeds the pace, pause and filler-word report
One microphone feeds two paths: the audio you can replay and download, and the transcript the delivery report is built from.

Delivery check: pace, pauses and filler words

A voice memo tells you what you said. It doesn't tell you how it came out, and that is the part you usually can't judge while you are talking. So every take you save here also gets a short report underneath it: what share of the clip you were really talking, how many pauses ran longer than 0.7 seconds, where the longest one was, and your speaking pace in words per minute.

Pace is the transcript's word count divided by your talking time, with the pauses taken out. Everyday conversation tends to land somewhere around 120 to 170 words per minute. If you rehearse a talk and the number says 190, that is the reason your friend said you sounded rushed. If it says 95, you are probably reading carefully, which is fine for dictation and a bit sleepy for a presentation. The pause count catches the opposite habit too: a ten-minute take with 40 pauses over 0.7 seconds is a lot of hunting for words. In English the report also lists possible filler words from the transcript, such as "like", "you know" and "basically".

To use it, record a rehearsal of at least 20 seconds, press Stop, and read the box that appears under the saved take. Then record it again with one change (slower, or fewer filler words) and compare the numbers.

The limits are real. Pace and fillers depend on the live transcript, so they only show up in browsers that support it, and not for Japanese, Chinese or Korean, where counting words by spaces means nothing. Chrome's recogniser often quietly drops "um" and "uh" before you ever see them, so the filler count is a minimum, and "like" is sometimes just a normal word. Talk time and pauses come straight from the audio, so those still work without any transcript. Background noise that stays loud can hide short pauses, because the page can't tell it from speech.


Checking pace and pauses in a 2 minute 30 second talk

Say you rehearse a 2 minute 30 second talk and save the take. The delivery check shows Talking 80% of clip, Pauses over 0.7 s: 3, Longest pause 1.1 s at 0:42, Pace 200 wpm. The talking figure is 120 seconds out of 150. The transcript has 400 words, and 400 divided by 2 minutes of actual talking gives 200. The note under it says that is fast and listeners will start to lose words, which matches the "you went too quick" comment you got last time.

The three pauses over 0.7 seconds are the clue: you hardly stopped. Record it again, this time breathing at the end of each point. The new take is 170 seconds long with 135 seconds of talking (79%), 9 pauses over 0.7 s and 360 words. That works out to 360 divided by 2.25 minutes, or 160 wpm, which the page calls inside the range most conversation falls in. The clip is longer, but the pace is easier to follow.

The opposite reading is not always a problem. Dictate a letter slowly and the page may say 95 wpm is on the slow side. For dictation or a careful read-aloud, that is fine. And if the filler line lists "like" three times, check the transcript first, because sometimes it was just a normal word.


What is this tool?

A dictaphone was a compact tape recorder that professionals and creatives relied on, especially in the final decades of the 20th century, to record voice notes for themselves. Smartphones made the physical device rare, but we still love that little cassette recorder — so here it is online, with one big upgrade the original never had: it transcribes your words into text live, right as you speak. If you just want a straightforward recording without the transcript or retro styling, our Voice Recorder is the simpler option.

Click REC and start talking. Your words appear on screen as you go, and the cassette's reels turn while it records. Click the square Stop button when you're done, and that recording is saved to your list below with both its audio and transcript, ready to play back, rename, or download.


What could you use this dictaphone for?

Interview recording

Get both the audio file and a text transcript at once, saving the hassle of transcribing everything by hand afterward.

Lectures and classes

If writing everything down is hard to keep up with, let the transcript capture it while you focus on listening.

Notes to yourself

Keep this page bookmarked or on your home screen, and record an idea the moment it hits you — easier to keep safe than a sticky note.

Commentary and voiceovers

Record your thoughts on something as it happens, then layer the audio over video footage afterward.

Language practice

Pick your target language from the dropdown, speak, and see how accurately the transcript catches your pronunciation.

Meeting minutes

Record the discussion and let the transcript do the note-taking, so everyone can stay present and engaged instead.

Diction practice

Rehearse a presentation or performance and review the transcript afterward to spot filler words like "um" and "like."

Family memoirs

Sit down with a relative and record their stories, building an audio archive worth keeping for years.

Bedtime stories

Record yourself reading a favorite story so it's ready to play whenever you can't be there in person.


How the live transcription actually works

Transcription runs on your browser's built-in Web Speech API, which streams short chunks of your audio to your browser vendor's speech-recognition service for conversion to text, then returns the words to this page as you speak. The audio itself is never saved or made available to us — only the text result reaches this page. Chrome and Edge support this well; Firefox and Safari have inconsistent or missing support, which is why the recording still works everywhere even when the live transcript doesn't appear.


Troubleshooting the transcript

No words are appearing at all

Check which browser you're using first — this is the most common cause by far. Live transcription needs Chrome or Edge; Firefox and Safari either don't support the underlying API or support it too inconsistently to rely on. Your audio recording will still work fine regardless. If you're on Chrome and still getting nothing, double-check microphone permission via Chrome's own help page before assuming the transcription itself is broken. Our Mic Test is a faster way to confirm the microphone itself is actually working before you dig any further into the transcription.

The transcript is full of mistakes

Accuracy depends heavily on microphone quality, background noise, and speaking pace. Moving to a quieter room, speaking a little slower and more clearly, and double-checking the language dropdown matches what you're actually speaking all noticeably improve results.

Transcription stopped partway through a long recording

Speech recognition sessions occasionally time out during silences. This tool automatically restarts listening when that happens, but a very long pause mid-recording can still show up as a gap in the transcript even though the audio recording itself continues uninterrupted.


How to use the online dictaphone

  1. Choose a language.

    Pick the language you'll be speaking, for accurate live transcription.

  2. Click REC.

    Allow microphone access when your browser asks.

  3. Speak.

    Your words appear on screen as you talk, and the cassette reels turn while it records.

  4. Click Stop.

    Your recording is saved to the list below with its audio and transcript.

  5. Play, rename, or download.

    Each saved recording can be played back, renamed, downloaded as audio, or removed.


Frequently asked questions

What is the Delivery check under each recording?

Once you stop, the page decodes your audio on your own device and reports how much of the clip you were actually talking, how many pauses ran past 0.7 seconds, the longest one, and your pace in words per minute using the live transcript. In English it also lists possible filler words. Pace and fillers need a transcript, so they only appear in browsers that support live transcription. Speech recognition tends to tidy up "um" and "uh", so the filler count is a minimum.

What is a dictaphone?

A dictaphone was a compact tape recorder that professionals used, especially in the late 20th century, to record voice notes for themselves. This tool brings that idea online, adding live speech-to-text transcription the original hardware never had.

How does the live transcription work?

It uses your browser's built-in speech recognition to convert your speech to text in real time as you talk. This works best in Chrome and Edge; some browsers don't support it, in which case the audio recording still works normally.

Is my recording uploaded anywhere?

This page never uploads your recording: the audio stays in your browser's memory, and closing or refreshing the page clears it, so download anything you want to keep. The live transcript is the exception to know about. It comes from your browser's own speech-recognition service, and Chrome and Edge normally process that audio on the vendor's servers. If that matters, record without relying on the transcript.

Can I record more than one voice memo?

Yes. Every time you click Stop, that recording is added to the list below, and you can immediately start a new one. Rename, play back, download, or delete each one independently.

What languages does the transcription support?

The language dropdown covers a range of major languages and regional English accents. Choose the one that matches how you'll be speaking before you hit record for the most accurate results.

Why isn't the transcript matching what I said?

Speech recognition accuracy depends on microphone quality, background noise, accent, and speaking pace. Speaking clearly at a moderate pace in a quiet room gives noticeably better results than a noisy environment.

Can I download my recording?

Yes, each saved recording has a download button that saves the audio file to your device.

Does this work on mobile?

Recording works on most modern mobile browsers. Live transcription support varies more by mobile browser than on desktop, so if words aren't appearing, the audio recording itself will still work.


More tools on ProDeviceTest

Testing your setup rather than recording a memo? These run the same way — instant, no download, right in your browser.

  • Voice Recorder — a simpler recorder for quick audio clips and playback tests
  • Mic Test — check that your microphone is picking up sound correctly
  • Tone Generator — play a precise test tone at any frequency
  • Sound Test — check your left and right speaker channels
  • All Audio Tools — the full category