What these tools actually cost you
Five downloaders, measured the same way on the same day.
By using this site you confirm you have the right to download the file you request.
What was measured, and how
Every "best downloader" list is somebody being paid. So rather than rank anything, this measures one thing that is not a matter of opinion: what each site costs your browser to open.
Each home page was loaded once in a headless Chromium at 1280 pixels wide, given seven seconds to settle, and every network request it made was counted. Weight is the encoded bytes over the wire. A third-party host is any domain outside the site’s own. Measured 9 November 2026 — anyone can re-run it and get their own numbers.
| Site | Requests | Page weight | Third-party hosts |
|---|---|---|---|
| downlo.net | 4 | 11 KB | 0 |
| cobalt.tools | 117 | 298 KB | 3 |
| ssstik.io | 57 | 842 KB | 11 |
| savefrom.net | 98 | 1,261 KB | 5 |
| snaptik.app | 79 | 1,450 KB | 36 |
What the numbers mean on a phone
A megabyte and a half before you have pasted anything is roughly five seconds on a slow connection and a visible chunk of a metered plan. It is also thirty-six other companies being told you are there.
The eleven kilobytes here are the page, the stylesheet and about five kilobytes of JavaScript — enough to read the box and draw the list of formats. There is no framework because there is nothing for one to do.
Where the others are better
cobalt is genuinely good and it would be dishonest to pretend otherwise. It is open source, it carries no advertising, and you can read exactly what it does — which nobody else on this list can say, including us, since this is not open source.
Its weight comes from being an application rather than a page, and its three third-party hosts are Cloudflare, not ad networks. If what you want is a clean tool from people you can audit, use it.
What it does not have is explaining. Its home page is 155 words and its sitemap lists 21 pages; this site has 295 and tells you what each platform will and will not give up.
The thing the measurements do not show
savefrom withholds 1080p unless you install its browser extension — the quality exists, it is held back as leverage. Its extension has been flagged as adware by more than one tracker.
y2mate, for years the most-used of these, was shut down along with eleven related sites by the IFPI in October 2025.
cobalt’s public API now requires a token, so the open endpoint that made it easy to build on is gone. That is their right and it is also a change worth knowing about.
None of that is visible in a page-weight table, which is the honest limit of this comparison.
What we do not claim
Not that this is the best tool. That depends on what you are downloading, and for some of it the answer is the platform’s own button — which the page on what this cannot do says plainly.
Not that the numbers will hold. Sites change; this one will be re-measured rather than left to go stale, and the date above is when it was taken.
Frequently asked questions
Why measure page weight rather than download speed?
Because download speed depends on your connection and the source platform, so it measures them rather than the tool. Page weight is the tool.
Can I check these numbers?
Yes — open each site with the network panel of your browser’s developer tools and read the request count and transferred size.
Are you not just promoting yourselves?
The table has us first because of one measurement with a stated method, and the section above says where cobalt beats us. A comparison that never concedes anything is an advertisement.