The clip-pipeline tool skill — choose the source, let AI find candidates, inspect and rank honestly, polish the top few, ship on a drip. The agent plans and QA-checks; the human reviews every clip; WoopSocial publishes. (Ships with ; pairs with the craft skill.)

OpusClip genuinely compresses 3–5 hours of manual clipping into minutes of processing plus review — its ClipAnything engine is multimodal (visuals + audio sentiment + facial expressions), so it finds moments even in wordless content, and natural-language moment search pulls one moment from a 2-hour VOD without scrubbing. But the top-1% operator holds two facts the marketing won't volunteer. (1) The Virality Score is triage, not truth: the only credible independent test found ~40% of generated clips get discarded and the score regularly mispredicts real performance — low scorers blow up, high scorers flop. Use it to rank where your editing time goes; never as an auto-post threshold or a guarantee ("85% faster" is a self-published number with no method — the ~40% discard is the figure you can plan around). (2) The human gate is non-negotiable: every clip passes three questions — does it stand alone, does it represent the speaker fairly (the interview meaning-spine applies to clips), does it serve the audience — before anything ships. And the billing mechanic that catches everyone: 1 credit = 1 minute of SOURCE video, regardless of clips out — trim before you upload.

(Depth: .)

2026 OpusClip: ~10M users / ~172M clips; ClipAnything (multimodal + moment search + negative prompts); speaker diarization layouts; ReframeAnything tracking; ~97% caption accuracy vendor-claimed across 20+ languages (QA anyway); AI B-roll; XML export to Premiere/Resolve (Pro). The score: 0–99 across Hook/Flow/Engagement/Trend — independently found to mispredict; 40% discard (BIGVU test, attributed); no independent benchmarks. Credits: 1 = 1 source minute; +1 per direct X post; failures reported eating credits; projects reported vanishing after subscription ends — download promptly. Tiers (≈, conflict — verify in-app): Free = evaluation only (watermark, 3-day expiry, no score/editor); Starter ≈ $15/150 min; Pro ≈ $29 ($14.50 annual)/300 min = the production tier; Business = custom, API reported Business-only; editor gating conflicts across sources. Reliability: reported processing hangs, no public status page/SLA — build slack. Trustpilot 4.0 with ~22% one-star (attribute). Full detail: . The weekly loop, per-clip QA, credit-math worksheet, score recalibration, and two worked examples: .

opus-clip (this) = the OpusClip-specific pipeline · captions-and-clipping = the general clipping craft (read first; covers the multi-tool layer — goes deep on this tool) · descript = the long-form talk edit (master there; its Underlord clip flags overlap) · capcut = short-form styling/polish beyond the built-in editor · short-form-video-script = the retention craft the QA applies · tiktok-video-publishing / instagram-reels-publishing / youtube-shorts = platform publish specifics.

Reads first: captions-and-clipping + the source skill + brand-profile/design-and-templates. Pulls the master from: descript (the edited episode), livestream-and-realtime (VODs), webinars. Feeds: capcut (polish), the platform publishing skills, content-recycling, scheduling-and-queue. Publishes via: approved exports → scheduling-and-queue → WoopSocial. Tool file:

tools/integrations/opus-clip.md

. Measure with: native + analytics-and-reporting on per-clip retention vs score (recalibrate monthly) — never fabricated.

A clip pipeline that treats OpusClip as a first-pass finder: source chosen and trimmed for the credit mechanic (1 credit = 1 source minute), candidates generated with genre/negative prompts set, the batch triaged by score as ranking only with a planned ~40% discard, every clip passed through the human three-question gate (stands alone / fair to the speaker with context intact / serves the audience), only the top 3–5 edited (hook line, caption accuracy on names/jargon, reframe drift, boundaries) with one brand template across the batch and auto-music off in favor of licensed/native audio, the approved set drip-scheduled and published via WoopSocial with exports downloaded promptly, and real retention tracked against the scores monthly to recalibrate; tier/editor/API gating verified in-app before subscribing (API automation claimed only at the reported Business tier; no unreviewed auto-posting anywhere); no score treated as truth, no out-of-context clips, no fabricated tiers/capabilities/ metrics; and correctly distinguished from captions-and-clipping, descript, capcut, and the platform publishing skills.