Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when video content needs to be extracted as text: pasted YouTube links or IDs, requests to transcribe, summarize, quote, translate, convert video to text, or extract information from video content. Also use when a user shares a video URL without explanation and wants to know what it says. Not for uploads or account management.
.claude/skills/bilal140202-video-transcript/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -7% | 0% |
Extract transcripts from videos via TranscriptAPI.com.
If $TRANSCRIPT_API_KEY is not set, read references/auth-setup.md and follow the instructions there to get and store the key.
Every request needs two headers:
Bearer $TRANSCRIPT_API_KEYHermesAgent/0.11.0, ClaudeCode/1.0). Version is optional — agent name alone is fine. Do not omit this header or send a bare default — Cloudflare will return a 403 (error code 1010) and block the request.bashcurl -s "https://transcriptapi.com/api/v2/youtube/transcript\ ?video_url=VIDEO_URL&format=text&include_timestamp=true&send_metadata=true" \ -H "Authorization: Bearer $TRANSCRIPT_API_KEY" \ -H "User-Agent: YourAgent/1.0"
| Param | Required | Default | Values | | ------------------- | -------- | ------- | -------------------------------------- | | video_url | yes | — | YouTube URL or 11-char video ID | | format | no | json | json (structured), text (readable) | | include_timestamp | no | true | true, false | | send_metadata | no | false | true, false |
Accepted URL formats:
https://www.youtube.com/watch?v=VIDEO_IDhttps://youtu.be/VIDEO_IDhttps://youtube.com/shorts/VIDEO_IDdQw4w9WgXcQResponse (format=text&send_metadata=true):
json{ "video_id": "dQw4w9WgXcQ", "language": "en", "transcript": "[00:00:18] We're no strangers to love\n[00:00:21] You know the rules...", "metadata": { "title": "Rick Astley - Never Gonna Give You Up", "author_name": "Rick Astley", "author_url": "https://www.youtube.com/@RickAstley", "thumbnail_url": "https://i.ytimg.com/vi/dQw4w9WgXcQ/maxresdefault.jpg" } }
Response (format=json):
json{ "video_id": "dQw4w9WgXcQ", "language": "en", "transcript": [ { "text": "We're no strangers to love", "start": 18.0, "duration": 3.5 }, { "text": "You know the rules and so do I", "start": 21.5, "duration": 2.8 } ] }
format=json when you need precise timestamps for quoting specific moments.send_metadata=true to get video title and channel for context.| Code | Meaning | Action | | -------- | ---------------- | ---------------------------------------------- | | 401 | Bad API key | Check key or re-setup | | 402 | No credits | Top up at transcriptapi.com/billing | | 403/1010 | Cloudflare block | Add or fix User-Agent header | | 404 | No transcript | Video may not have captions enabled | | 408 | Timeout | Retry once after 2s |
1 credit per successful request. Errors don't consume credits. Free tier: 100 credits, 300 req/min.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 10,869 | 5,265 | -52% | 1 | 1 | 0% | 2,733 | 1,384 | -49% | 0 | 0 | — |
case-16 | pass→pass | 4,315 | 1,345 | -69% | 1 | 1 | 0% | 597 | 1,135 | +90% | 0 | 0 | — |
case-02 | fail→fail | 6,918 | 5,613 | -19% | 1 | 1 | 0% | 1,441 | 1,319 | -8% | 0 | 0 | — |
case-03 | pass→pass | 7,053 | 3,597 | -49% | 1 | 1 | 0% | 1,499 | 1,498 | -0% | 0 | 0 | — |
case-04 | pass→pass | 9,853 | 3,444 | -65% | 1 | 1 | 0% | 2,011 | 1,704 | -15% | 0 | 0 | — |
case-05 | fail→pass | 9,863 | 3,875 | -61% | 1 | 1 | 0% | 1,955 | 1,871 | -4% | 0 | 0 | — |
case-17 | fail→pass | 7,550 | 1,795 | -76% | 1 | 1 | 0% | 1,355 | 1,317 | -3% | 0 | 0 | — |
case-06 | fail→pass | 10,307 | 2,315 | -78% | 1 | 1 | 0% | 1,832 | 1,480 | -19% | 0 | 0 | — |
case-07 | fail→pass | 11,601 | 2,682 | -77% | 1 | 1 | 0% | 2,233 | 1,542 | -31% | 0 | 0 | — |
case-08 | pass→pass | 7,957 | 1,913 | -76% | 1 | 1 | 0% | 1,597 | 1,378 | -14% | 0 | 0 | — |
case-09 | pass→pass | 7,789 | 1,759 | -77% | 1 | 1 | 0% | 1,472 | 1,352 | -8% | 0 | 0 | — |
case-10 | pass→pass | 5,980 | 1,254 | -79% | 1 | 1 | 0% | 1,100 | 1,222 | +11% | 0 | 0 | — |
case-11 | pass→pass | 8,263 | 1,443 | -83% | 1 | 1 | 0% | 1,530 | 1,189 | -22% | 0 | 0 | — |
case-12 | fail→pass | 7,446 | 1,806 | -76% | 1 | 1 | 0% | 1,320 | 1,231 | -7% | 0 | 0 | — |
case-13 | pass→pass | 5,189 | 2,001 | -61% | 1 | 1 | 0% | 1,007 | 1,287 | +28% | 0 | 0 | — |
case-14 | fail→pass | 8,827 | 1,956 | -78% | 1 | 1 | 0% | 1,622 | 1,288 | -21% | 0 | 0 | — |
case-15 | pass→pass | 6,955 | 1,404 | -80% | 1 | 1 | 0% | 1,165 | 1,211 | +4% | 0 | 0 | — |
case-18 | pass→pass | 4,422 | 2,367 | -46% | 1 | 1 | 0% | 829 | 1,410 | +70% | 0 | 0 | — |
case-19 | pass→pass | 7,840 | 1,412 | -82% | 1 | 1 | 0% | 1,387 | 1,196 | -14% | 0 | 0 | — |
case-20 | fail→pass | 10,685 | 1,274 | -88% | 1 | 1 | 0% | 1,962 | 1,168 | -40% | 0 | 0 | — |
case-21 | pass→pass | 7,972 | 1,715 | -78% | 1 | 1 | 0% | 1,340 | 1,285 | -4% | 0 | 0 | — |
case-22 | pass→pass | 12,647 | 7,679 | -39% | 1 | 1 | 0% | 2,597 | 2,662 | +3% | 0 | 0 | — |
case-23 | pass→pass | 7,011 | 5,073 | -28% | 1 | 1 | 0% | 1,406 | 2,064 | +47% | 0 | 0 | — |
case-24 | pass→pass | 7,766 | 5,551 | -29% | 1 | 1 | 0% | 1,524 | 2,094 | +37% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted, and 22 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +29 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.