YouTube Video Uploader

YouTube Video Uploader: publish a file from a URL to your own channel through the official API

Point it at a public link to your video file, give it a title and your own Google OAuth credentials, and it uploads to your channel with the description, tags, hashtags, privacy, schedule and thumbnail set in the same run. It is built to be the publish step at the end of a pipeline: render somewhere, drop the file on a CDN, hand this actor the URL.

The awkward part is Google's, not ours, and it catches almost everyone. A new Google Cloud project starts in Testing mode, where your refresh token expires after seven days, and an unverified project has every upload forced to private whatever you asked for. Publish your consent screen and get the project verified before your first public upload. The row reports privacyOverridden: true when YouTube overruled you, so at least you find out from the output rather than from your channel.

📺 What YouTube Video Uploader does

It downloads your file, signs in with your refresh token and calls videos.insert, then thumbnails.set if you gave it a thumbnail. Nothing about your channel is kept after the run.

Privacy fails closed. It defaults to private, so nothing goes public unless you ask for it. If you set a schedule on a public or unlisted video the actor moves it to private for you and says so in the log, because YouTube will not schedule anything else. Afterwards it reads back the privacy YouTube actually stored, which is how privacyOverridden gets filled in.

There is a second browser method that drives YouTube Studio with exported cookies instead. It was built for running on your own machine, and it ignores tags, publishAt, categoryId and madeForKids. On Apify, use the API method.

📥 What you give it

{
  "videoUrl": "https://cdn.example.com/clips/episode-12.mp4",
  "title": "How the Roman aqueducts actually worked",
  "description": "A short breakdown of gravity-fed water systems.",
  "tags": ["history", "engineering", "rome"],
  "privacyStatus": "private",
  "publishAt": "2026-06-15T14:00:00Z",
  "categoryId": "27"
}

📤 What you get back

One row per run. There is no real sample row published here yet, so rather than invent one, these are the fields and what each of them means:

🧾 Reading the output

That split is deliberate: a preview cannot end up looking like a delivery, and it cannot bill like one either.

A failure does not write a row. The run itself ends as failed, with a message naming the fix rather than a stack trace:

On the browser method a failed step also saves a screenshot and the page HTML into the run's key-value store, and names them in the error.

▶️ How to run it

1. Create an OAuth client in Google Cloud Console with the YouTube Data API enabled, and generate a refresh token carrying the youtube.upload scope. 2. Move the consent screen to In production, and request the compliance audit if you intend to publish anything public. 3. Open YouTube Video Uploader, paste the three credential fields, and put your file's link in videoUrl. 4. Fill in title, pick a categoryId, and leave privacyStatus on private for the first run. 5. Click Start, then read videoId, watchUrl and privacyOverridden off the row.

Run it with no credentials at all and you get the preview row instead, so you can see the shape before setting any of this up.

💰 How much does it cost?

$0.05 per uploaded video. Flat on every Apify plan, no volume tiers.

A run that fails before the upload completes, and a preview run, are not charged that fee. It lands on uploads that reached your channel, not on attempts.

Your Google quota is separate and it is Google's, not ours: videos.insert costs about 100 units against a default of 10,000 a day, so roughly 100 uploads a day before you need an increase.

💡 What people use it for

• The last step of a render pipeline, so the file never waits in a bucket for somebody to upload it.
• Scheduling a batch: one run per video, each with its own publishAt, all private until their

moment.

• Re-publishing a back catalogue with consistent titles, tags and category from your own data.
• Posting to a second channel on the same account, named by channelId.

🚧 What it does not do

• It cannot lift Google's private-video override. An unverified project forces every upload

private and only Google's compliance audit changes that. The row tells you it happened.

• It does not manage your OAuth for you. There is no shared app here: the credentials, the

project and the quota are yours.

• The browser method ignores tags, publishAt, categoryId and madeForKids, and it always

declares the video as not made for kids. If your video is child-directed, use the API method, or set the declaration yourself in Studio.

• The API method collapses your description onto one line. Line breaks do not survive the

upload.

• One video per run. There is no queue and no playlist step.
• It does not wait for HD processing. A video looks soft for a while after upload and sharpens

on its own.

• It does not check what is behind your videoUrl beyond the file type, so point it at links you

control.

🧭 Which YouTube actor do you need?

❓ Questions people ask

Why did my public upload come out private? An unverified Google Cloud project. YouTube forces private and does not tell you at upload time. Request the compliance audit, and watch privacyOverridden in the meantime.

My refresh token stopped working after a week. The consent screen is still in Testing mode, where Google expires tokens after seven days. Move it to In production.

Do I need to give you my Google account? No. You create the OAuth client, you hold the refresh token, and it is marked secret in the input.

How many can I upload a day? Google's quota decides, not this actor. The default works out at roughly 100 uploads a day.

Can I see the output before setting up OAuth? Yes. Run it with no credentials, or tick demoMode, and the preview row lands in the run's key-value store.

🆘 If something breaks

Open the Issues tab on the actor page with the run id. The failure message usually names the fix already, and on the browser method the saved screenshot in the key-value store shows where it stopped.