Google Drive as a media host: bulk uploads with public links

A bulk scheduler that imports posts by CSV or API doesn't take file uploads — it fetches every video from a public URL at import time (Metricool's CSV and its API both work this way). So a batch of rendered videos needs a host that is: free at real batch sizes, scriptable, and boring. Google Drive is all three.

The pattern is always the same, whatever tool sits downstream:

1. Upload the files.
2. Set them to "anyone with the link" can view.
3. Hand the share links to the scheduler.

By hand, for small batches

1. In drive.google.com, create a folder for the batch and drag all files in at once.
2. Right-click the folder → Share → General access → "Anyone with the link", role Viewer. Files inherit the folder's access, so one setting covers the whole batch.
3. Per file: right-click → Copy link. That link goes in the CSV.

The link format

The link you copy looks like this:

https://drive.google.com/file/d/FILE_ID/view?usp=sharing

That's the one to use. It technically points at Drive's viewer page, not the raw bytes — but import backends recognize the pattern and resolve it to the file server-side. We've pushed rendered videos north of 70 MB through Metricool's CSV import and its normalize endpoint on exactly this format.

There is also a direct-download form:

https://drive.google.com/uc?export=download&id=FILE_ID

Scripted, for real batches

Uploading thirty files by hand once is fine; every week, it isn't. The Drive API does the whole pattern in a few lines per file (Python, google-api-python-client):

from googleapiclient.http import MediaFileUpload

media = MediaFileUpload(path, resumable=True)
file = service.files().create(
    body={"name": name, "parents": [folder_id]},
    media_body=media,
    fields="id",
).execute()

service.permissions().create(
    fileId=file["id"],
    body={"type": "anyone", "role": "reader"},
).execute()

link = f"https://drive.google.com/file/d/{file['id']}/view?usp=sharing"

The permission call is per file and explicit: scripted uploads don't pass through the folder-sharing UI, so grant type: "anyone", role: "reader" right after each upload.

The one-time chore is the OAuth client (a credentials.json from Google Cloud Console and one consent click). It's the same setup VidVertex uses — the walkthrough is here: Google Drive setup.

Click to enlarge

×

Housekeeping

• Storage math: the free tier is 15 GB, shared with Gmail and Photos. Rendered vertical videos run tens of MB each, so a few hundred fit — but a weekly pipeline will want an occasional cleanup or a paid tier.
• Know when the copy stops being needed. Metricool re-hosts every file on its own servers — at normalize on the API route, at import on the CSV route — so the Drive copy is dead weight from then on. For a scheduler whose behavior you haven't verified, keep the files until the posts have actually published.
• One folder per run keeps cleanup honest: when a batch is through, the folder can go as a unit.
• Deleting means deleting. A trashed file still counts against the quota until the trash is emptied.

Where VidVertex fits

Both of VidVertex's Metricool routes run this exact chain automatically at the end of an upload run: every rendered brand-and-platform variant goes to your Google Drive (resumable, drive.file scope, share permission set per file). In Metricool (Drive + CSV) the ready-to-import bulk CSV lands next to it with the links already in place; in Metricool API the link is normalized and the app deletes its own Drive copy right after (upload modes). Google Drive is never a publishing path here — it is the transfer vehicle, nothing else.

Connect Drive once — the OAuth-client walkthrough — and the hosting step, cleanup included, disappears from your week.