Video & Animation

Read public YouTube data with the YouTube Data API v3 and an API key: search videos, channels and playlists; get video details and statistics; resolve a channel from a @handle or URL and list its uploads and playlists; list playlist items; read comments; get trending videos and categories. Use when a task needs facts or numbers from YouTube (research, competitor or channel reports, finding videos, collecting comments). Don't use to upload, edit, delete, like or comment (those need OAuth, not an API key), to read private or unlisted data or YouTube Analytics (watch time, revenue), or to download videos or captions.

把任意主题做成视频，按内容形态自动选择渲染后端。黑底 MG 概念科普（Remotion）、连续公式推导证明（Manim）、真终端操作教学（asciinema，V2）、产品宣传短视频（Hyperframes，V3）。内含渲染器无关的公共层（调研规范、口播文案原则、TTS 配音时间轴、分镜令牌系统、读帧 QC 脚本）、后端路由判据树、可插拔渲染后端、多 agent 并行构建协议与 QC 判据。Turn any topic into a narrated video — routes by content shape to the right renderer: motion-graphics concept explainers (Remotion), continuous formula derivations and proofs (Manim), real terminal recordings for command-line teaching (asciinema, V2), and product promos (Hyperframes, V3). Includes a renderer-agnostic core (research brief, narration principles, TTS timeline builder, storyboard tokens, frame-reading QC scripts), a routing decision tree, pluggable backends, and multi-agent build/QC protocols.

Edit video and audio with local FFmpeg from natural-language requests: cut, trim, join, resize/reframe (9:16, 1:1), speed change, captions and subtitles (SRT/ASS, animated, karaoke), logos and text overlays, lower-thirds and titles, silence removal, multicam and external-mic sync, loudness normalisation, HDR/Dolby Vision to SDR, LUTs, background music with ducking, platform exports (YouTube, Reels, TikTok, X), compliance checks, scene detection and highlight reels, contact sheets to inspect results, and whole-edit project files. Use this skill whenever the user mentions a video or audio file (mp4, mov, mkv, wav, m4a), footage, a clip, captions, subtitles, a reel or short, YouTube/Instagram/TikTok delivery, LUFS, sync, transcoding, ffmpeg, or asks to make something "60 seconds", "vertical", "louder", "captioned" — even when they do not say "edit". Python 3.10+ standard library only (3.9 deprecated), no cloud, no API keys.

Agent 2 of the video editing pipeline. Finds and removes spoken mistakes — stutters, repeated words, false starts, and retakes (re-recorded lines) — from a talking-head recording. Use after cut-silences, when asked to cut mistakes, remove stutters / repeats / filler restarts, clean up flubs, or keep the best take. Works review-gated: it proposes every cut with context and a reason for approval, then renders only the approved cuts via ffmpeg. Requires a word-level transcript.

A-to-Z video editing from a plain-language brief, for creators (Facebook, Instagram, YouTube, TikTok), ads/OVC and any everyday edit. Analyses the footage, asks the user option-based questions, then cuts, reframes, colour grades, captions (incl. Bangla), adds music, SFX, voice-over, motion graphics (Remotion) and 3D titles (Blender), cleans voice (Demucs), transcribes (Whisper), finds music by description (CLAP) and exports per platform. Use whenever a video, clip, reel, short, ad, slideshow or footage edit is requested. Not for generating new footage from text.

Agent 1 of the video editing pipeline. Removes silences and dead air from a talking-head recording. Use when asked to cut silences, trim pauses, remove dead air / gaps, or tighten the pacing of a raw video, given a word-level transcript. Produces an edit list (EDL), a re-timed transcript for downstream agents, and optionally the cut video via ffmpeg. Does NOT cut mistakes, repeats, or false starts — that is the cut-mistakes agent.

「苏格拉没有底」视频工坊：两种已验证风格：【漫画口播】（混知式彩色漫画 + 用户自录口播逐字对齐，竖屏 1080×1920）与【像素卡点】（像素暖光故事、无口播、文字对白，画面卡在配乐乐句上，4:3 1440×1080）。做本账号任何视频时加载，先按 §1 选风格，再读对应 STYLE.md。

Make a clean 45–90 second product demo video for any software product, rendered from code to MP4 with music. Airy light canvas, short centred lines the camera pushes into, the product's own UI rebuilt and shown up close as it is used (typing, results, tables, clicks), integrations, and a logo close. Use when someone wants a product demo, launch video, feature walkthrough or "how it works" video for an app, SaaS or tool.

Make a 12–20 second high-energy motion-graphics promo for any product, rendered from code to MP4 with synthesised music. Fast beat-locked scene changes, kinetic typography, a shape that carries across every transition, a HUD frame. Use when someone wants a hype reel, teaser, launch sting or showreel-style promo for a product, app or brand.

Build Japanese lyric subtitles for a music video the way Japanese readers actually read them: break lines at 文節 (bunsetsu) boundaries, reveal each phrase at its MEASURED sung onset (not by averaging syllable counts), write 漢字仮名交じり rather than all-kana, and verify every character of the line also appears in the background typography. Use when a task involves Japanese song lyrics on screen, karaoke/phrase timing, 字幕断句, ルビ/縦書き, lyric typography, or 'the subtitles look wrong to a Japanese viewer'. Ships two scripts: a measured-onset extractor (pyin + energy gating, with the phrase-end false-positive fix) and a coverage checker.

Revise narration, design professional timecoded storyboards, and produce 2D/3D animation for evidence-led short videos. Use when a creator provides a script, production brief, source material, or reference film and needs any of these three stages or a finished non-presenter explainer. 适用于口播稿调整、专业分镜、2D/3D动画和逐镜对标制作；不用于简单裁剪或单张封面。

Make DeepSeek (or any model) produce motion-design videos that match what Claude Opus 5.5 makes from the same prompt, for any request in any style — brand films, showreels, kinetic typography, explainers and data stories, product and UI promos, character stories and music videos, abstract loops, title sequences. It reads the prompt like Opus, derives an art direction from it, plans before coding, builds a Remotion (React + SVG) project on a tested starter with generated sound, and reviews its own work with a five-juror critique. Triggers on 动效, 动态设计, 视频, 短片, 动画, 复刻 opus 效果, 扩写提示词, Remotion, motion graphics, showreel, MV.

Make vertical social media reels (Instagram, TikTok, Shorts) from the user's own footage with the reel-kit project (Remotion + ffmpeg). Use when the user asks to make, edit, re-cut or render a reel or short promo video from their videos, for any topic (yoga, aikido, classes, events).

Verifies an output of the studio (video, image, audio, 3D model) before accepting it: extracts frames and LOOKS at them, checks the format with ffprobe, measures the loudness and runs `make verify`. Use it after every render, after changing the audio or after touching the brief.

Move a Webflow site off Webflow onto Astro + Tailwind, statically built and served by Cloudflare Pages, page for page: same URLs, titles, meta, OG, canonicals, JSON-LD, sitemap, robots, redirects and analytics, with forms, CMS content, video and the domain moved too, so the Webflow site plan can be cancelled. Use whenever someone wants to leave Webflow, cut Webflow costs, "export my Webflow site", rebuild a Webflow site in code, move a Webflow site to Cloudflare/Astro/static hosting, replace Webflow forms or CMS (with Astro content collections, Sanity or another headless CMS), or asks what breaks when you leave Webflow. Any kind of site: company or product sites, portfolios, shops (Webflow Ecommerce), restaurants and local businesses, nonprofits, publications, events, membership and multi-language sites. Also for Webflow-to-Next or other static stacks: the inventory, parity, forms, CMS and DNS steps are the same. With the seo-aeo-geo skill installed, the move also improves search and AI visibility.

The voxel career-film style. A blocky 3D version of the person walks, without ever vanishing, through one scene per role. Their outfit changes as they walk, each scene rises out of the floor ahead of them, and a dark banner tracks year, company and role. Use it when someone wants their career, a journey or a "how I got here" told as a short 3D film for LinkedIn or a portfolio, says "make my career film", "turn my CV into a video", "voxel film", or wants a new chapter added to an existing one. It runs on film-engine (render, audio, finish) and must pass film-audit before anything is sent.

Concept, design, animate and deliver motion-graphics videos built in code (an HTML page where every frame is render(t), recorded frame-exact to mp4) at a level that rivals After Effects work, in twelve styles (cinematic product film, clean light product film, atmospheric glass, narrated explainer, a day in the life, character-led, hero shot, sizzle, card showcase, walkthrough, kinetic type, data story). Use this whenever someone wants a video of any kind: a product launch or feature explainer, a portfolio or personal-brand video, a staff or customer training video, a how-to walkthrough, a change/update notice, a teaser or social short, an animated data story, a demo, or "turn this into a video" / "make a video like that one". Also use it for the pieces around a video: concept and storyboard, voice-over script for AI voice tools, fitting animation to a recorded voice, music mix, captions, thumbnail, and the post that goes with it. Works in any language.

Build OpenMAIC-inspired, analogy-first interactive knowledge-chain webpages that make complex questions understandable through familiar scenes, explicit metaphor mappings, sandboxed hands-on labs, targeted video support, and sourced phenomenon-to-boundary explanations. Use for deep topic explainers and "Why First" learning pages. Do not use for simple factual answers, API documentation, or ordinary marketing pages.

Build and deploy a cinematic scroll-driven website for any business or idea. Use when the user asks to build a website, landing page, cinematic site, scroll site, one-page site, or hero-video site for any business, product, brand, place, portfolio, or concept (plumber, lawyer, hotel, SaaS, coffee shop, design studio, anything). Also use when the user wants to set up the tools this workflow needs, including the Higgsfield CLI for AI image and video, the manual alternative of prompts for ChatGPT (images) and Google Flow (video), the Hostinger connector for hosting and deploy, ffmpeg, or Node.js.

Product Launch: turn a product (a URL, an app the user can record, screenshots or a brief) into the kind of 20–60 s launch video SaaS teams post on X: kinetic headlines and the real product UI in floating windows, cut to a music grid, ending on the name and a call to action. Real UI only, never mocked. Works with whatever video tool the project uses (Hyperframes, Remotion, Motion Canvas, an editor); Hyperframes when starting fresh. A free local script makes a music bed authored to the cut. Every version is reviewed in FrameJam. Use for a product launch, feature announcement, "introducing X" video, promo for a website or app, or a landing-page hero video. Not for a founder talking to camera (founder-launch) or a short mystery teaser (launch-teaser).

Direct, build and render premium cinematic motion design from code — launch films, product and feature videos, looping feature animations for landing pages and social, logo stings and reveals, UI walkthroughs, app demos, kinetic type, teasers, trailers, intros and promo clips. Use this whenever the user wants any video, animation, motion graphic or animated demo made with Remotion, HyperFrames, HTML/GSAP or React, or asks to storyboard, time, score, render, polish or critique one — even if they only say "make a video about X", "animate this feature" or "we need a launch clip". It supplies concept and copy discipline, type and colour taste that defers to a supplied brand, a beat-locked timeline that drives both picture and a code-made soundtrack, choreography and transition craft, real product UI in motion, film-grade finishing (true motion blur, no banding, BT.709, sync-checked mux) and a measured self-critique loop on contact sheets and renders.

Premium brand-kit image generation skill for creating high-end brand-guidelines boards, logo systems, identity decks, and visual-world presentations. Trained for minimalist, cinematic, editorial, dark-tech, luxury, cultural, security, gaming, developer-tool, and consumer-app brand systems. Optimized for intentional logo concepting, refined composition, sparse typography, strong symbolic meaning, premium mockups, art-directed imagery, and flexible grid layouts.

Translate, repair, or selectively polish English-to-Hebrew SRT subtitles. Use for Hebrew subtitle translation, synchronized SRT repair, RTL punctuation, SDH cleanup, character-aware gender, recurring series terminology, legacy Hebrew encodings, translator/site credit removal, full bilingual editorial review, and batch or ZIP subtitle QA.

Replace a video's original English vocals with Chinese voiceover on an identity timeline — the video keeps its original speed and duration, each cue's Chinese audio is fitted into its original window. Use when the user wants to dub a video into Chinese — mentions 中配 / 配音 / 中文配音 / 换原声, or has a cooked bilingual video and wants a second Chinese-narrated release, or another skill (e.g. video-cooking) hands off "video is done with subtitles, add a Chinese dub.