X (Twitter) Video Downloader

X (Twitter) Video Downloader: save the video from a public post as MP4, or its sound as MP3

Paste links to public X or Twitter posts, one per line. Each video is saved into the run's key-value store as a real file, and the dataset row that goes with it carries a direct download link, the title, the author, the resolution and the exact file size.

The part worth knowing first: this only reaches posts anyone can open signed out. A protected account, a post behind a login, or a deleted one comes back as a sentence in the run's ERRORS record and costs nothing.

🎬 What X Video Downloader does

You give it post links. For each one it reads the post, picks the format, checks the file size before a single media byte moves, downloads it, checks the file actually decodes, stores it, and only then lists it in the dataset.

X publishes the same picture twice, as one ready-made MP4 and as a streaming version split into picture and sound. This takes the ready-made MP4, so the file arrives whole with its audio already in it and its size known in advance. Nothing is stitched together afterwards.

With Audio only on you get a 128 kbit/s MP3 of the same post instead.

No login, no cookies, nothing to paste in besides the link.

📥 What you give it

{
  "urls": [
    "https://x.com/historyinmemes/status/1790637656616943991",
    "https://twitter.com/someone/status/1234567890123456789"
  ],
  "quality": "1080",
  "audioOnly": false,
  "maxMegabytes": 500
}

A post with several videos gives you the first one. End the link with /video/2 to ask for the second instead.

Duplicates are dropped, and a link that points somewhere else is refused with a note telling you which actor takes it. A t.co short link is refused too: open it, and paste the address it lands on.

File size is bounded by the run's memory. A run can hold a file of roughly 70% of the memory it was given, so about 716 MB at the default 1,024 MB. If you want something bigger than that, raise the run's memory as well as maxMegabytes.

📤 What you get back

Two things per link: the file in the run's key-value store, and a row in the dataset pointing at it.

A real row, from run 4kPIloPIN7IwX9npC, with the signed link trimmed:

{
  "url": "https://x.com/historyinmemes/status/1790637656616943991",
  "host": "x",
  "title": "Historic Vids - One of the most intense moments in history",
  "author": "Historic Vids",
  "durationSeconds": 15.56,
  "resolution": "728x720",
  "format": "mp4",
  "fileSizeBytes": 1404023,
  "billedMegabytes": 2,
  "mediaKey": "x-1790637589910654976.mp4",
  "mediaUrl": "https://api.apify.com/v2/key-value-stores/E2Boy2HNCFDNWIzez/records/x-...?signature=..."
}

🧾 Reading the output

A failure never reaches the dataset. Every row in there is a file that exists and was stored.

Links that did not work go into a key-value record called ERRORS, each as {url, error} with the reason written in plain words. It is saved as each failure happens, so a run that runs out of time still tells you what it managed and what it did not.

The sentences you will actually see:

If every link in a run fails, the run ends as failed and the status message points at ERRORS.

▶️ How to run it

1. Open X (Twitter) Video Downloader and click Try for free. 2. Paste your post links into X post links, one per line. 3. Pick a Max quality, or tick Audio only (MP3). 4. Set Max megabytes per video if you want a tighter cap than 500. 5. Click Start, then take mediaUrl from each dataset row, or open the run's Storage tab.

💰 How much does it cost?

$0.00045 for every megabyte delivered, counted in whole megabytes and the same on every Apify plan. The 15-second clip in the example above counted as 2 MB.

You are billed only after a file is stored and listed, so a link that fails, a post that is private or gone, a link to another site, and a video skipped for being over maxMegabytes all add no megabytes. A run that delivers no files has none to bill.

Higher quality means a bigger file, and a bigger file costs more. audioOnly is much smaller than video on the same post.

💡 What people use it for

• Keeping a copy of a clip before the post that carries it disappears.
• Pulling the audio out of an interview or a speech to transcribe it somewhere else.
• Feeding mediaUrl into an editor, a captioner or an uploader as the next step in a pipeline.
• Archiving a set of posts at a fixed quality so file sizes stay predictable.

🚧 What it does not do

• Public posts only. Protected accounts and anything behind a login fail with a sentence saying so.
• No photos, no GIF-only posts, no Spaces and no broadcasts.
• The first video of a multi-video post unless you ask for another with /video/<n>.
• No post text, likes, replies or metrics. The title carries the post's text and that is all.

For the post as data, use a scraper instead.

• No t.co links. They are refused with an instruction rather than followed.
• No subtitles or captions.
• Files live in the run's key-value store, which is not permanent. Copy anything you want to

keep, and treat mediaUrl as a link to fetch now rather than store for later.

• One file per link, so a run's total is bounded by what the run can hold in memory.

🧭 Which downloader do you need?

❓ Questions people ask

Do I need an X account or cookies? No. Nothing to paste in and nothing that expires.

Where is my file? In the run's key-value store. The fastest way to it is mediaUrl on the dataset row; the Storage tab of the run shows the same records under mediaKey.

Can I download a whole account's videos? Not from here. This takes links you supply. Use a scraper to collect the post links first, then feed them in.

Why is my video smaller than 1080p? Because the post does not have a 1080p version. You get the closest size X published.

What happens if one link is dead? It goes to ERRORS with a reason, adds nothing to your bill, and the run carries on with the rest.

Is downloading these videos legal? The post being public does not make the video yours. Copyright still applies, and so does data-protection law where a person is identifiable. Have a reason for what you are collecting. Apify's write-up on scraping and the law is a fair start, and we are not lawyers.

🆘 If something breaks

Open the Issues tab on the actor page. Send the run ID and the exact post link, and paste the ERRORS record from the run's key-value store: it carries the sentence for every link that failed.