Uploading Videos to YouTube API

There are two valid upload patterns depending on your runtime and file size constraints.

Approach A: Google API Client (googleapis) Upload

Good for standard Node services with local file access.

const response = await youtube.videos.insert({
  part: ['snippet', 'status', 'localizations'],
  requestBody,
  media: { body: createReadStream(videoPath) },
});

Useful features:

• direct metadata + media call
• easy scheduled publishing (publishAt)
• straightforward retry wrappers

Approach B: Resumable Streaming Upload

Good for Worker/edge pipelines, large files, or R2 stream sources.

Session Init

const response = await fetch(
  'https://www.googleapis.com/upload/youtube/v3/videos?uploadType=resumable&part=snippet,status,contentDetails',
  { method: 'POST', headers, body: JSON.stringify(metadata) }
);
const uploadUrl = response.headers.get('Location');

Chunk Upload

• Use 256KB aligned chunks.
• Send Content-Range each PUT.
• On 308, parse Range to reconcile acknowledged bytes.

if (response.status === 308) {
  const range = response.headers.get('Range');
  const match = range?.match(/bytes=0-(\d+)/);
  bytesUploaded = match ? Number(match[1]) + 1 : bytesUploaded + chunk.length;
}

Metadata Defaults (Common News Pipeline)

• category 25
• defaultLanguage='ja'
• containsSyntheticMedia=true
• privacy policy-driven (public when clean, private otherwise)

Thumbnail Upload (Recommended)

After main upload:

• upload JPEG thumbnail via thumbnails/set
• enforce 2MB limit with resize/quality fallback
• treat thumbnail failure as warning if main video is already uploaded

Which Approach to Choose

• Use Approach A for simple app servers and local files.
• Use Approach B for distributed pipelines, stream sources, and strict memory environments.