DEV Community

Mobeen Ahmed
Mobeen Ahmed

Posted on

Why I Left Douyin as "Coming Soon" Instead of Wiring a Data-Only API

Every platform page on my site has the same promise: paste a link, get the file. TikTok works, Instagram works, Facebook and YouTube work. Douyin, the Chinese version of TikTok, still says "coming soon". That is a deliberate choice, and it comes down to what the available APIs actually hand you.

What a downloader page actually needs

A downloader is not a search box. When someone pastes a link, my code has to resolve it to a real video file — a URL the browser can save as an MP4. That is the whole job. Everything else is decoration.

For TikTok, Instagram, Facebook and YouTube, the APIs I use return exactly that: a direct file URL, sometimes with a choice of quality. My server checks the file, proxies it through, and the visitor gets a download. Simple, honest, and it either works or it tells you plainly that it did not.

What the Douyin APIs actually return

When I went looking for a Douyin API, the options I found did not do that. What they return is metadata — titles, descriptions, thumbnails — plus stream playlists. A playlist is not a file. It is a list of small chunks the video player stitches together while you watch. You cannot hand a playlist to a browser and say "save this as an MP4".

I could have wired one of those APIs up anyway. The page would look finished. It would show the video title, the thumbnail, maybe even a preview player. And the download button would either do nothing useful or quietly hand the visitor a broken file. That is the part nobody sees in the demo but everybody notices in real use.

There is a second problem with playlists. Even if I stitched the chunks together on my server, the result would be a re-encoded file, not the original video. It would cost my server time and bandwidth on every download, it would fail on long videos, and the quality would be whatever my re-encode decided. That is a different product from the rest of the site, wearing the same label.

The "coming soon" label is the honest version

So the Douyin page on ramrom.io stays a placeholder, and it says plainly that it is coming soon. I would rather have a page that admits a feature does not exist yet than a page that pretends it does.

This is a rule I keep coming back to on this project. A tool site lives or dies on whether the button does what it says. One fake feature — one download button that does not download — costs more trust than ten missing features. Visitors do not remember the ten platforms you support. They remember the one that wasted their time.

What would change my mind

The door is not closed. If an API appears that returns a genuine downloadable file for Douyin links, the same way my other providers do, I will wire it in the same afternoon. The test is simple: paste a real Douyin link, and check whether what comes back is a file or a playlist. Until an API passes that test, the honest answer is "not yet".

That is the same test I apply to every new platform request. The question is never "can I put a page up for it?" — putting a page up takes an hour. The question is "can the visitor actually download the video?" If the answer is no, the page waits.

Building a small tool site teaches you this quickly. The hard part is not the code. The hard part is refusing to ship something that looks done when it is not.

Top comments (0)