What a downloader does, step by step

Five things happen, in this order

The word downloader covers a lot of very different programs, so it is worth naming what this one is. It is a reader and a fetcher. It opens a page, works out which media files that page is serving, tells you what it found, and then collects the one you chose.

Each of the five steps below is a plain mechanical operation. Knowing the sequence explains why the wait is where it is, why the rows say what they say, and why the file you end up with looks the way it does.

One: the page is opened and read

Our servers request the address you pasted and receive the same page a browser receives. Where the page forwards elsewhere, those hops are followed until the request lands on the page carrying the video.

The page is then processed rather than merely stored. Most sites assemble their player with scripts, so the reading includes running that assembly to the point where the player knows which files it intends to play.

Two: the published media files are identified

A page that plays video is pointing at one or more media files on a server somewhere. Those pointers are what the reading is looking for. They sit in the page markup, in the player's configuration, or in the manifest a streaming player uses to list its quality options.

Everything found this way is something the publisher put there for visitors to receive. A page holding no such pointer produces nothing, which is why a channel listing, a search result, or a licence-protected catalogue returns an empty list.

Three: what exists is measured and listed

Each file that was found is checked for its size, and the resolution it was encoded at is read from its properties. Those two figures become the row you see. A list of five rows means the publisher produced five separate encodings of that video.

This is where the tool stops and waits for you. It has not decided anything on your behalf; it has reported an inventory. An audio row appears in that inventory when the source stores its sound as a separate file.

• A row is one real file, so two rows at the same resolution differ in bitrate and therefore in size.
• The tallest row available is the tallest the publisher produced, since nothing in the inventory was created here.
• The size on the row is the size that will be written to your storage.

Four: the file you chose is fetched

Pressing Download file names one entry in that inventory. Our servers request that specific file from the host holding it and receive its bytes, in the same way your browser would if it were playing the video.

Nothing about the content is altered on the way through. The bytes handed on are the bytes the publisher's server sent.

Five: your browser writes it to storage

The file is handed to your browser, and your browser does what it does with every download: writes it into the folder it is configured to use and lists it in its own downloads view. That is why the progress appears there rather than on this page, and why closing this tab partway leaves you with a fragment.

Once it has finished, the copy belongs to your device. Prepared files on our side expire on a short timer; the one on your disk does not.

The saved file is a copy of the published file

A common picture of how this works has the video playing somewhere and being captured, frame by frame, into a new file. That is screen recording, and it is a different operation with different results: it takes as long as the video runs, and it produces a fresh encoding that is a generation away from the original.

What happens here is a transfer of an existing file. It takes as long as the megabytes take rather than as long as the video lasts, which is why a two-hour recording can arrive in minutes. The picture quality is exactly what the publisher encoded, because it is the publisher's encoding.

The same reasoning explains a question that comes up constantly. A page whose top offering is 720p yields a 720p file, and no setting on this site changes that, because the tool collects what exists rather than producing something new.

What that leaves you responsible for

Reading a page and fetching a published file says nothing about whether you are entitled to keep the result. The licence on somebody's work is unchanged by where the file is stored.

Using it costs nothing, within three downloads each day and a 150 MB limit on any one file. Pro lifts that daily figure, lifts the file size limit, and clears the advertising away. Neither tier asks you to register, and the video you keep sits on your own device.

Related pages

• Using the tool on a video now
• Containers and codecs in plain words
• How resolution and bitrate set the size
• Why no conversion step exists here
• What your browser handles at the end