The work behind an online download

The address arrives at a machine, not at a person

When you press Analyze, the text in the box is sent to one of our servers. Your own device stops working at that moment and waits for an answer.

That division is why a five-year-old phone and a new desktop behave the same way here. The reading, the measuring, and the fetching all happen on hardware you are borrowing for a few seconds.

Online download means servers open the page, measure published media, and prepare the row you chose so your browser can write the file. Time varies with how hard that page is to read and how large the file is.

Nothing useful stays queued under your name after you finish. The durable result is the file in your download folder.

Opening the page the way a visitor's browser would

The first piece of work is a request to the site you named. The server asks for that page, receives the same HTML a browser receives, and runs the scripts the page depends on to put its player together.

This is where an address that names a collection, a page behind a sign-in, or a video that has been taken down stops. The server is answered honestly by the site, and the answer contains no video.

Listing what the page publishes, then measuring it

The second piece of work reads the assembled page for the media files it points at. A page frequently publishes the same video several times over at different resolutions and bitrates, and each of those is a separate file with its own address.

Each candidate is then asked for its size. That request returns the length of the file without transferring it, which is how a figure in megabytes appears beside a row you have not downloaded yet. The list you see is the result of that measuring, sorted and handed back to your browser.

Preparing the single file you chose

Nothing is fetched until you press Download file. That press names one row, and only that row is acted on.

The server requests that file from the source, receives it, and holds it ready. The wait here follows the size of the file and the speed of the source, and it happens between our server and the site rather than over your own connection.

Handing the finished file to your browser

The last piece of work is the transfer to you. Your browser is given the prepared file and writes it into your downloads location, and this is the only stage your own connection is involved in.

The tab has to stay open across that transfer, because the browser tab is the thing receiving the bytes. A prepared file sits on the short expiry timer, so start the transfer rather than leaving the ready row alone for an hour.

Why one address takes longer than another

The reading stage dominates the wait on pages that build their player with a lot of script, because the server has to run that script before there is anything to list. A page that publishes plain media in its markup is read in a fraction of that time.

The preparing stage is dominated by the file itself. A ten-megabyte clip and a hundred-megabyte recording go through identical steps, and the second one occupies the server for ten times as long at the same source speed.

What the server keeps when you are finished

The prepared copy expires on its short timer and the server stops holding it. The copy that lasts is the one your browser wrote to your device.

That is also why a ready row cannot be forwarded to somebody else as a link. Send them the saved file, or send them the address of the page you started from and let them run the same four pieces of work.

Keep this tab open until the browser finishes writing the file for download video link online. A prepared file only waits a short time before it expires from the temporary store.

Related pages

• What happens between Analyze and the file arriving
• Why the pause after Analyze varies
• How the transfer behaves inside your browser
• Running this with nothing installed on your device