Study a Recording — Guide

The problem this solves

A lot of what people learn from now arrives as a recording — a talk, a lecture, a broadcast, an interview, a class someone sent you. And a recording is the hardest kind of study material there is. You cannot skim it. You cannot search it. You cannot write in the margin. When something lands that you want to remember, the only thing to do is scrub backwards afterwards and hope you find it again.

Reading material got all the tools and listening material got none. This is the part of Comment Coach that closes that gap: bring a recording in, and it becomes something you can mark, search, quote, comment on and come back to — with every note tied to the exact second it belongs to.

Your recording is never stored on our servers, and neither is its transcript — both live in this browser. What reaches our server is an address: this recording, this part, second 842, never the words that were spoken. Transcribing is the one exception, and it is worth being exact about: the audio is cut into parts, each part is sent to the transcription service, read, and dropped. Nothing is written to disk on the way through, and the text comes back here.

Getting it in

From the Library, choose Import and pick the file.
Or drop it on the uploader from your dashboard.

Audio or video — MP3, M4A, WAV, MP4 and the rest. A video is treated as its soundtrack, so a lecture you were sent as an MP4 works the same as one you were sent as an MP3.

Listen, properly

A waveform rather than a bare progress bar, so you can see where the talking is and where the pauses are. Skip back and forward fifteen seconds. Speed from 0.75× to 2×. Loop a passage until it sticks — which is the single most useful thing there is if you are working in a second language.

It plays offline. Once the file is in, nothing about listening to it needs the network.

Mark the moment as it happens

This is the part that does not exist anywhere else. While it is playing, one tap drops a mark at the second you are on — 👍 good, ❓ ask about this, 🔁 use this, ✋ disagree. It does not pause. It does not ask you to type. You keep listening.

Or press Mark to stop and write a note — the mark is saved the instant you press it, before the note is written, so a thought you never finished is still a thought you kept.

Every mark shows twice: as a tick on the waveform, and as a list underneath with its timecode. Tap a row to jump there. Walk forwards and backwards through them without looking at the screen.

Turn it into text

Transcribe reads the whole recording and writes it out. Every paragraph keeps the second it was spoken at, which is what makes the rest work: click a paragraph to hear it, and the audio and the text never disagree about where you are.

Marks you made before transcribing are not lost. They are re-filed into whichever part now contains their second.

This is the one step that costs anything. Everything above it — playing, marking, clipping, keeping — is free and needs no key.

It arrives in parts

A long recording is split into parts with titles, so the contents page reads like a table of contents rather than a wall. If the file carries its own chapter marks — most podcasts and many published talks do — those are used exactly as they were written, because the person who made it already knew where the divisions were.

Prefer it whole? Read it all on the contents page puts every part back into one continuous document, with each heading still a link into its own part.

Comment on any paragraph

Once it is text, it behaves like every other document here. Select a sentence and write what you think. Ask Hoot about it. Get coaching on what you wrote. Put it in a talk outline alongside points from somewhere else entirely.

Your comment is anchored to the timecode, so it stays attached to the right moment even if the recording is transcribed again.

Keep the part, not the hour

Save just this part cuts that section out as its own audio file you can keep or send. Keep only these parts throws the rest away — useful when ten minutes of a ninety-minute session is the bit you actually wanted.

Stop anywhere, carry on anywhere

Where you got to is remembered, and it follows your account. Listen on the train, open it on the laptop that evening, and it starts where you stopped.

If you are studying in a second language

A transcription can produce the other language beside the original, so the two sit in parallel columns at the same moments. With loop and speed on the same screen, a passage you could not follow at full speed can be heard slowly, read in both languages, and marked — without leaving the page.

Nothing here needs the recording to be yours. But everything you make from it is: your marks, your notes, your comments. Your copy of the recording is never published, shared, or made searchable by anyone else, and we never go and fetch a recording on your behalf.

Where things are

The recording — in this browser's own storage, and nowhere on our servers. Release it any time and keep the transcript. If you transcribe, the audio passes through the transcription service and is not kept there either.
The transcript — on this device only. It is included in your library backup, which is the only copy that exists, so take one.
Your comments and marks — on your account, as addresses, so they reach your other devices.

Tips

Mark first, think later. A one-tap mark while listening costs nothing and you can add the words afterwards.
If a file has chapters, you skip the guessing entirely — and it costs less, because nothing is sent to be structured.
Very long recordings are limited on a phone and generous on a desktop. If it refuses, the message says what to do about it.
A transcript cannot be re-made from nothing. Back your library up.