Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Remove pauses, filler words (um, uh), and dead air from raw YouTube recordings via the Tubeify API. Use when the user wants to edit a video, clean up audio, trim silences, or polish a raw recording for YouTube.
.claude/skills/davepoon-tubeify/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 9 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 116% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -43% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -25% | 0% |
Submit a raw recording URL to the Tubeify API and get back a polished, trimmed video with pauses, filler words, and dead air removed automatically.
bashcurl -c session.txt -X POST https://tubeify.xyz/index.php \ -d "wallet=<WALLET_ADDRESS>"
Response on success:
json{ "status": "ok", "session": "active" }
If the response contains "status": "error", check the wallet address and retry.
bashcurl -b session.txt -X POST https://tubeify.xyz/process.php \ -d "video_url=<URL>" \ -d "remove_pauses=true" \ -d "remove_fillers=true"
Parameters:
video_url (required) — direct URL to the raw video fileremove_pauses — remove silent gaps and dead air (default: true)remove_fillers — remove filler words like "um", "uh", "like" (default: true)Response on success:
json{ "status": "queued", "job_id": "abc123" }
bashcurl -b session.txt https://tubeify.xyz/status.php
Poll every 15 seconds. Terminal states:
| status | Meaning | Action | |------------|--------------------------------|-------------------------------| | queued | Waiting in queue | Keep polling | | processing | Actively editing | Keep polling | | complete | Finished — download ready | Read download_url from body | | failed | Processing error | Check error field, retry |
Complete response example:
json{ "status": "complete", "download_url": "https://tubeify.xyz/dl/abc123.mp4" }
Failed response example:
json{ "status": "failed", "error": "Unsupported video format" }
bashcurl -o edited_video.mp4 "<download_url>"
| Variable | Description | |----------|-------------| | TUBEIFY_WALLET | Ethereum wallet address for authentication |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 6,584 | 5,787 | -12% | 1 | 1 | 0% | 1,139 | 1,087 | -5% | 0 | 0 | — |
case-02 | fail→pass | 6,302 | 9,914 | +57% | 1 | 1 | 0% | 1,254 | 2,710 | +116% | 0 | 0 | — |
case-03 | fail→fail | 4,275 | 5,619 | +31% | 1 | 1 | 0% | 612 | 879 | +44% | 0 | 0 | — |
case-04 | fail→pass | 12,290 | 3,155 | -74% | 1 | 1 | 0% | 2,109 | 1,212 | -43% | 0 | 0 | — |
case-10 | fail→pass | 10,239 | 2,344 | -77% | 1 | 1 | 0% | 1,918 | 926 | -52% | 0 | 0 | — |
case-05 | pass→pass | 8,666 | 3,135 | -64% | 1 | 1 | 0% | 1,753 | 1,114 | -36% | 0 | 0 | — |
case-06 | fail→pass | 7,027 | 1,246 | -82% | 1 | 1 | 0% | 982 | 746 | -24% | 0 | 0 | — |
case-07 | fail→pass | 7,625 | 2,120 | -72% | 1 | 1 | 0% | 1,276 | 962 | -25% | 0 | 0 | — |
case-08 | fail→pass | 13,025 | 3,203 | -75% | 1 | 1 | 0% | 1,999 | 1,275 | -36% | 0 | 0 | — |
case-09 | fail→pass | 10,047 | 2,180 | -78% | 1 | 1 | 0% | 1,667 | 999 | -40% | 0 | 0 | — |
case-11 | fail→fail | 5,934 | 3,909 | -34% | 1 | 1 | 0% | 1,019 | 872 | -14% | 0 | 0 | — |
case-12 | pass→pass | 9,340 | 3,867 | -59% | 1 | 1 | 0% | 1,447 | 1,049 | -28% | 0 | 0 | — |
case-13 | fail→pass | 10,064 | 1,399 | -86% | 1 | 1 | 0% | 1,818 | 724 | -60% | 0 | 0 | — |
case-14 | pass→pass | 7,682 | 1,867 | -76% | 1 | 1 | 0% | 1,137 | 745 | -34% | 0 | 0 | — |
case-19 | pass→pass | 8,546 | 2,370 | -72% | 1 | 1 | 0% | 1,483 | 983 | -34% | 0 | 0 | — |
case-15 | fail→pass | 7,569 | 1,796 | -76% | 1 | 1 | 0% | 1,297 | 891 | -31% | 0 | 0 | — |
case-16 | pass→pass | 4,896 | 1,466 | -70% | 1 | 1 | 0% | 765 | 800 | +5% | 0 | 0 | — |
case-17 | pass→pass | 3,467 | 2,400 | -31% | 1 | 1 | 0% | 543 | 886 | +63% | 0 | 0 | — |
case-18 | fail→pass | 10,736 | 1,565 | -85% | 1 | 1 | 0% | 1,589 | 831 | -48% | 0 | 0 | — |
case-20 | fail→pass | 9,619 | 4,305 | -55% | 1 | 1 | 0% | 1,862 | 1,442 | -23% | 0 | 0 | — |
case-21 | pass→pass | 19,526 | 10,548 | -46% | 1 | 1 | 0% | 3,541 | 2,472 | -30% | 0 | 0 | — |
case-22 | pass→pass | 12,225 | 3,673 | -70% | 1 | 1 | 0% | 1,930 | 1,227 | -36% | 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. 22 cases were attempted, and 20 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 +50 percentage points is the difference between those two pass rates over the 20 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.