Knowledge Bank

4 min read

The transcript downloader, and reading a competitor's opening

Pull any public video's captions, cleaned but never reworded, and read the first 150 words as structure.

Not sure if Plan is your weak phase? The free 3-minute workflow audit will tell you.

Flat illustration of a video frame with a caption bar.
PlanLinked task: Scout the competition · step 3 of the 47

Scout the competition is step 3 of the 47, and most people do it by watching five videos and forming an impression. Reading the transcript instead turns an impression into something you can mark up. The transcript downloader pulls the caption track off any public video and hands it back in a form you can actually work in.

Paste the URL
Pick the caption track
Toggle the cleanup
Take the .txt or the .srt

Why raw captions are unusable

Auto-generated captions arrive in lowercase, split mid-sentence, littered with bracketed noise like [Music], and carrying the same line twice where the speaker paused. Nobody can read a hook structure out of that. You end up skimming, which is the thing you were trying to stop doing.

The cleanup pass fixes the mechanical problems. It strips the noise cues, restores sentence casing, spaces the punctuation properly, drops the duplicated lines, and fixes contractions that are unambiguous in English.

Nothing gets reworded. Every change is a deterministic rule, so the words stay exactly as the speaker said them. Toggle any rule off and the file rebuilds instantly.

That last part matters more than it sounds. A transcript that has been quietly tidied by a language model is no longer evidence of what anybody said, and evidence is the entire reason you pulled it.

The three files, and what each is for

FileUse it for
Cleaned .srt, with timingsEditing to timecode, and re-uploading a corrected caption track
Original .srt, exactly as YouTube serves itWhen you need to see what was really there, untouched
Plain .txt, no timestampsWriting from. Paste it into a doc and mark it up

Reading a competitor's opening

This is the use that changes how you plan. Pull the transcript of a video in your niche that clearly worked, then read only the first 150 words.

You will see the shape immediately: how many sentences before the promise lands, whether they name the outcome or the problem first, where the first specific detail arrives. That shape is very hard to hear while you are watching, because the delivery carries you past it. On the page it is obvious.

Do this across three or four videos and the pattern in your niche stops being a matter of opinion. Take it into drafting your own three hooks.

Your own back catalogue is the other half

One video already contains the Shorts script, the newsletter paragraph, the chapter markers, and most of a blog post. All of it is sitting in a caption track you have never opened.

Before a repurposing run, pull the transcripts for everything you plan to mine. Working from text is far faster than scrubbing a timeline hunting for the bit you half remember, and the timings in the .srt tell you exactly where to cut.

What it will and will not do

It reads any public video or Short that has captions, manual or auto-generated, in any language the video carries. If a video has no caption track at all, there is nothing to pull, and no tool can invent one.

Three videos are free before it asks for anything. One email unlocks it permanently, with no card and no per-video cap.

Auto-captions get names wrong

Worth holding in mind if you are pulling a transcript to quote a figure or a product name. Auto-captions mishear proper nouns constantly, and they do it confidently, so a mangled brand name arrives looking exactly like a correct one.

The tidying pass does not touch this, because guessing at what a word was meant to be is the rewording it deliberately refuses to do. If a number or a name is going into your own script, watch that section of the video and check it with your ears.

This is also the argument for correcting the caption track on your own uploads. The transcript people read, the one search engines index, and the one anybody repurposing your work will pull is the one sitting on the video, mistakes and all.

The mistake worth avoiding

A transcript is raw material, and raw material is not a script. Pasting a competitor's transcript into a document and rewriting it sentence by sentence produces something that reads like a cover version and performs like one.

Read for structure, borrow nothing verbatim. The useful thing on the page is the order they put ideas in, not the sentences they used to say them.

Mark up where the promise lands, where the first proof arrives, and where they stop setting up and start delivering. Then close the file and write your own.

Keep reading

Once you have read a few openings, take what you found into how to write a YouTube hook. The transcript is also the fastest route into cutting Shorts from long-form, and if you are heading for the description instead, chapters are a four-minute job with the timings already in hand.