yt-dlp and Instagram: “login required”, empty results and what to use instead

The message bundles three different situations into one line: the post is unavailable, your IP hit a limit, or Instagram wants a session. On a datacenter address it is nearly always the third — the request never reaches the post, it hits the login wall first. yt-dlp cannot tell them apart from the outside, which is why its advice is “try cookies”.

$ yt-dlp https://www.instagram.com/reel/ABC123/
ERROR: [Instagram] ABC123: Requested content is not available,
rate-limit reached or login required. Use --cookies,
--cookies-from-browser, --username and --password, or --netrc

Exporting a browser session works the same day and stops working within days: Instagram rotates session tokens, ties them to the IP and device they came from, and flags accounts that fetch hundreds of posts from a server. The fix then becomes a rota of accounts and proxies to manage — which is a scraping operation, not a download command.

Before removing yt-dlp from our own pipeline we ran 13 public reels through it from a Hetzner address: 0 succeeded, and single calls ran up to 45 seconds before failing. Instagram's own embed endpoint did no better — 0 of 11. The links were fine; a home connection resolved them. It is the address, not the tool.

One GET, no cookies on your side, no browser. The resolving happens on a chain of sources that do not depend on an Instagram session, and the response has the same shape for a reel, a post, a carousel or a story. The median response time is measured live and shown on the front page.

curl -s "https://api.saveapi.org/v1/instagram" \
  -H "Authorization: Bearer sk_live_xxxx" \
  -G --data-urlencode "url=https://www.instagram.com/reel/ABC123/"

If your code called yt-dlp for the URL and then downloaded the file yourself, this is the same contract: pass the link, get back direct media URLs, fetch the bytes from the CDN. What you lose is yt-dlp's format selection; the response carries one rendition per Instagram media, so there is nothing to select.

import httpx

def resolve(link: str) -> list[str]:
    """yt-dlp o'rniga: havoladan to'g'ridan-to'g'ri media URL'lar."""
    r = httpx.get(
        "https://api.saveapi.org/v1/download",
        params={"url": link},
        headers={"Authorization": "Bearer sk_live_xxxx"},
        timeout=30,
    )
    data = r.json()
    if not data["success"]:
        raise RuntimeError(data["error"]["code"])
    return [m["url"] for m in data["medias"]]