YouTube API Quota Exceeded: Causes and Fixes

Every Google Cloud project gets a default YouTube Data API v3 quota of 10,000 units per day, reset at midnight Pacific time - there's no rollover and no way to trigger an early reset. Quota is tracked per project, not per API key, so generating a second key from the same project doesn't give you a second pool.

Every API call costs units, not requests - and the costs are uneven. Here's the full picture across the endpoints developers actually use:

At 100 units per search, a project can exhaust its entire daily quota after just 100 search calls - before a single caption is downloaded. This is the single most common way small projects hit the limit: they didn't make 10,000 requests, they made a few hundred expensive ones. By contrast, videos.list, channels.list, and playlistItems.list are cheap at 1 unit each - if you already know a video, channel, or playlist ID, those calls barely dent the daily budget.

1. Open the Google Cloud Console for the project tied to your API key.
2. Go to APIs & Services → YouTube Data API v3 → Quotas & System Limits.
3. This shows your current usage against Queries per day and the per-minute limits, updated with a short delay rather than in real time.

If you're already near the ceiling before you've made 10,000 requests, this view is how you confirm it's the expensive-call problem above, not a bug.

Both return a 403, and both get lumped together as "quota errors" in casual conversation, but they're different problems with different fixes:

• quotaExceeded (reason: dailyLimitExceeded) - your project used its full daily unit allocation. Nothing fixes this except waiting for the midnight Pacific reset, reducing per-call cost, or getting a quota increase approved.
• rateLimitExceeded - you're sending requests too fast in a short window, independent of your daily total. This is a pacing problem, not a budget problem - see our rate limit guide for how to handle it correctly with backoff.

1. Cache aggressively. Quota is charged per API call, not per user-facing request - if you re-fetch the same video's captions repeatedly, cache the result (even a simple in-memory TTL cache) instead of calling the API again for data that doesn't change minute to minute.
2. Avoid search.list where a direct ID lookup works. If you already have a video, channel, or playlist ID, the corresponding .list call is 100x cheaper than a search.
3. Batch ID lookups. videos.list accepts up to 50 comma-separated IDs per call for the same 1-unit cost as a single ID - fetching 50 videos one at a time instead of batched is 50x more expensive for no reason.
4. Request a quota increase once you have real production usage to justify it, via the YouTube API Services Audit and Quota Extension form linked from the Cloud Console's quota page. Expect a manual review, not instant approval.
5. Use a transcript API that doesn't run through your own YouTube quota. GetYouTubeTranscript's API fetches transcripts against its own infrastructure rather than proxying captions.download with your key, so there's no daily 10,000-unit ceiling to manage - only the plan's own request-rate limit.

cURL:

curl "https://getyoutubetranscript.com/api/v1/transcript?v=dQw4w9WgXcQ" \
  -H "Authorization: Bearer sk_live_..."

Node.js:

const res = await fetch(
  "https://getyoutubetranscript.com/api/v1/transcript?v=dQw4w9WgXcQ",
  { headers: { Authorization: `Bearer ${process.env.GETYOUTUBETRANSCRIPT_API_KEY}` } }
);
const { data } = await res.json();
console.log(data.transcript);

Python:

import os, requests

res = requests.get(
    "https://getyoutubetranscript.com/api/v1/transcript",
    params={"v": "dQw4w9WgXcQ"},
    headers={"Authorization": f"Bearer {os.environ['GETYOUTUBETRANSCRIPT_API_KEY']}"},
)
print(res.json()["data"]["transcript"])

This costs 1 credit against your GetYouTubeTranscript plan, not units against your Google Cloud project's YouTube Data API quota. Get 100 free credits, no card required, from the dashboard, or see the full API docs.