Video & Animation

Run the whole `manga` CLI (manga-anim) pipeline from Claude Code, on the Mac or through API keys (OpenRouter, fal, OpenAI, Google, xAI, Replicate, ElevenLabs, Anthropic) — reading video (speech bubble OCR, voices, captions, camera moves), AI animation of panels, a full anime episode generated from the manga, and a new manga (spin-off) drawn with the characters of an existing one. Use when the user wants to read, edit, dub or animate a manga, comic or webtoon, turn one into an anime, "from this page to that page", import a CBZ/PDF/EPUB, search MangaDex, fix a project's OCR, or invent a sequel or a new story with existing characters.

Fixt de eerste 3 seconden van een korte video (reel, TikTok, Short). Geef je openingszin, wie je kijker is en wat de video waarmaakt; je krijgt een diagnose, 3 nieuwe hooks in verschillende vormen, een check tegen 6 soorten kijkers en de beste keuze. Gebruik bij "fix mijn hook", "betere opening", "hook-meester", "waarom swipen ze weg", of als je een script of openingszin plakt.

Edit raw video footage inside Claude through the Higgsfield connector. Runs a clip through color grading, a subject cut-out (rotobrush), clean up (removing unwanted objects) and the /miracle offer for the free trial, with ready prompts for each step. Use when the user uploads or links a video clip and asks to grade it, make it look cinematic, cut out or mask a person or object, remove something from the shot, put text behind a subject, or asks about the /miracle commands, Higgsfield in Claude, or the Higgsfield free trial.

Make a narrated learning-explainer video (MP4) that teaches a concept, in one soft 3D clay look with a silent mascot, with its script checked by one agent before the voice is made. Use when the user asks for a video, an explainer, an animation or a visual walkthrough of an idea, such as "make a video explaining eigenvectors", "generate an explainer on I2C", "animate how attention works". Uses what the current conversation and project already show, so run it in the conversation where the topic came up.

AI video editing for talking-head recordings, run locally. Automatic smart cuts that remove long pauses, filler words and redone takes, interview and podcast editing, vertical videos for Reels, TikTok and Shorts, automatic captions/subtitles (SRT and burned-in), and an animated podcast / cartoon mode where a paper-cutout scene follows the narration above the speaker. Use when the user asks to edit a video automatically, cut pauses or silences, remove "um"/"uh", clean up retakes, edit an interview, add captions or subtitles, make a vertical video, or turn a recording into a cartoon or animated podcast.

Make animated story videos in 10 styles that do big numbers on Reels and TikTok: pixel art, PS2 style, comic art, lego style, claymation, paper cutout, retro anime, low poly 3D, risograph and miniature diorama. Agent writes the story, codes every scene, and renders an mp4 on the user's own computer, with no credits or image models. Use one style for a whole video or switch style every beat. Use when asked for an animated video, a story video, a video in any of these styles, "make it look like claymation / paper / pixel art", an animated ad for an app or SaaS, or when someone wants story copy for a video.

Two ready-made visual style templates for Remotion promo films (宣传片 / 形象片 / VCR), aimed at schools, colleges and institutions. 苹果风格 is Apple-style frosted glass in blue, white and a little yellow, with 优设标题黑 titles. 华章版 (Aurum Nocturne) is midnight lacquer with gold leaf. Includes a working engine template (scene kit, per-shot looks, transitions, photo frames, near-full-screen card rail, headline-number emphasis, text-as-mask opener), render, QA, master and replacement-segment scripts, and design principles. Use it whenever someone wants a promo film for a school, college, institute or company made in Remotion. Also use it for requests that mention 苹果风格 (formerly 磨砂版), 华章版, 磨砂风, 毛玻璃, 黑金, 苹果风, frosted glass or a dark-gold luxury look, for a new edition or restyle of an existing promo, and for replacement segments of one, even if the skill is not named.

Build coded prototypes that behave like the real product, share them for feedback, then turn them into polished walkthrough videos for case studies, decks and sites. Use when asked to prototype a flow, make a clickable demo, or make a product walkthrough or reel from a design.

Use when planning, prompting, generating, editing, routing, finishing, or QAing AI-created images, GIFs, videos, style frames, image-to-video clips, social ads, cinematic brand assets, motion studies, or multi-provider generative media workflows using Codex imagegen, GPT Image, Sora, Grok Imagine, Veo, Luma, Runway, Remotion, Playwright, ffmpeg, Figma, Canva, or Vercel.

Making a whole code-rendered video (every frame computed by a program): intro ads, hype ads, product tours, showreels, narrated explainers — audience, tone, sections, transitions, camera moves, export. Trigger on 「做一支產品介紹片/宣傳片/廣告」「酷炫、抓眼球的廣告」「做 showreel」「做旁白說明動畫」「影片怎麼分段、換段、轉場、運鏡」「這支片的基調/受眾」. NOT one UI animation's easing (→ motion-design), studying reference videos (→ case-library).

Use when making UGC-style video ads. Covers Reddit-pain scripts, the full avatar pipeline (FAL voice clone, Seedance walled garden, 480p + ffmpeg upscale, word-by-word verification, Mirage Tesseract captions/motion graphics, ~$5), human creators via Tribe/Sideshift, and disclosure rules.

Make and clone short-form and YouTube videos with Remotion and 2D animation. Built-in format is stick-figure "You Decide" story shorts (9:16 YouTube Shorts, TikTok, Instagram Reels) where viewers vote A or B and the next part follows the winning path. Also recreates ANY example video the user provides (explainers, top-N lists, story narrations, history/fact videos, quizzes, would-you-rather, skits, chat stories, kinetic text, whiteboard, horizontal long-form) by analyzing its cuts, pacing, palette, captions and sound, then rebuilding the format with original content. Covers scripts, stick-figure rig, sets, captions, natural TTS voices, copyright-free synthesized SFX and music, mixing and rendering in vertical, horizontal or square. Use for stickman/stick figure animation, "you decide" or choose-your-own-adventure shorts, the next part after a vote, faceless channel videos, or "make a video like this one".

Writes and rewrites video scripts with the open-loop cycle — Risk, Big Question, Misdirection, Re-hook — so every section ends with a new reason to keep watching. Use when drafting a script body, fixing a flat or predictable script, or patching drop-off points between sections. Persian-first, works in any language.

Turns lecture materials (usually a PDF of slides) into a single, fully narrated explainer MP4 video for self-study and exam prep – in a selectable language (same as the slides, German, English or custom), with an AI voice via OpenRouter MCP, highly readable graphics, slide images from the PDF, code/SQL/formulas, an intro (course, topic, professor), an optional outline with time ranges, a summary, optional self-test questions, subtitles and chapters. ALWAYS use this skill when someone uploads lecture slides, a script or university materials and wants a video, video summary, explainer, narrated presentation or "something to watch instead of attending" – even if the word "video" never appears. German examples - "mach mir ein Video aus den Folien", "erklär mir die Vorlesung als Clip", "mach mir was zum Lernen für die Klausur aus diesen Folien", "Video-Zusammenfassung der Vorlesung". English examples - "lecture summary video", "turn these slides into a narrated explainer".

Versi gratis One Line Motion (10 gaya: Kartun Flat, Papan Sains, Pixel 8-bit, Retro Neon, Panel Komik, Low-poly 3D, Clay, Sinematik, Stickman, Cat Air). Ubah satu kalimat ide jadi video motion graphic bersuara (MP4): explainer, Shorts/Reels/TikTok, intro, promo, iklan produk, kinetic type. Alur: brief → 3 konsep → naskah per adegan → suara → animasi Remotion → QA → render → paket upload. Pakai saat pengguna minta bikin video animasi, motion graphic, explainer, konten edukasi bergerak, atau merevisi video One Line Motion.

Writes and audits scroll-stopping video hooks (Reels, Shorts, YouTube, TikTok) using a six-block word system — subject, action, goal, contrast, plus optional proof and time. Use when writing the first line of a script, rewriting a weak hook, or analysing why a hook works. Persian-first, works in any language.

Structures and audits video scripts with the six-step curiosity ladder — stimulation, fascination, anticipation, validation, affection, revelation — so viewers stay to the end and come back. Use when planning a video's flow, diagnosing where viewers drop off, or checking a script before filming. Persian-first, works in any language.

Make Claude direct 3D videos like a 3D animator (camera, light, motion and sound), written frame-by-frame in code with Remotion + three.js. Use whenever someone asks for a 3D video, a 3D explainer, a product/object animation, a cinematic or "cool" 3D clip, a vertical reel/short in 3D, or wants an existing Remotion/three.js scene to look less flat and more cinematic.

A curated collection of 1000+ agent skills from official dev teams and the community, compatible with Claude Code, Codex, Gemini CLI, Cursor, and more.

اشتغل من الآيباد كأن عندك كمبيوتر قوي: موشن جرافك بالكود في دقايق، ورندر Blender بـ Cycles، ونصوص عربية على الفيديو، وتصميم صوت، وتجميع بـ ffmpeg، ومزرعة رندر من جلسات سحابية متوازية، كلها داخل جلسات Claude Code السحابية بدون PC. استخدمها كلما قال المستخدم ipad_code أو /ipad_code، أو قال إنه على آيباد أو جوال أو ما عنده كمبيوتر، أو طلب موشن جرافك أو مقطع Blender أو رندر ثقيل أو أنيميشن ثلاثي الأبعاد أو تسريع الرندر بأجهزة كثيرة، حتى لو ما ذكر اسم السكل. Use for any heavy creative/compute job (3D render, video, audio) run from an iPad through Claude Code cloud sessions.

电影解说配音、短剧解说配音、影视解说配音技能：解说旁白（narrator）加多个角色台词，各用不同音色，一条命令生成整段 AI 配音 WAV、SRT 字幕和逐句时间轴 manifest.json，按时间轴在剪映 / CapCut / Premiere 里对画面。音色可用账号已有角色或参考音频声音克隆，打造解说号固定声线；语气标签和语速可调；改一句稿只重配一句。适合电影解说号、电视剧解说、短剧解说、动漫解说、纪录片旁白。配音由 Miragea 开放 API 完成，本地只需 Python 3.9+ 标准库、无需显卡，可 --dry-run 离线试跑。适用于 Claude Code、Codex、OpenClaw、Cursor、WorkBuddy。Movie recap AI voiceover, multi-voice narration, voice cloning, SRT subtitles.

Turn any process, workflow, pipeline, decision loop or system into an animated flowchart explainer with a design made for that topic — a token travels the flow with real speed control, every check sends a query out to whatever answers it (a rules engine, a reviewer, a model) and the answer comes back, a loop-back plays out, and live read-outs show what got written and where time went. Use this whenever the user wants a flowchart, process, architecture or loop "animated", "made into a video/GIF", "explained like that diagram video", "做成流程圖/動畫", an agent or AI pipeline, approval flow, data pipeline, banking/IT workflow, onboarding or incident process visualized — even if they never say "skill" or "Motion".

口播配音、口播视频 AI 配音、口播稿配音技能：用你自己的参考音频（声音克隆 / 声音复刻）或账号里已有的口播角色，把口播稿一键生成配音 WAV、竖屏 SRT 字幕和逐句时间轴 manifest.json；说错一句只改一行重新生成，不用对着镜头反复录。适合抖音、视频号、小红书、快手、B站的知识口播、财经口播、情感口播、职场口播博主和 MCN 批量更新；可用语气标签和语速控制口播节奏。配音由 Miragea 开放 API 完成，本地只需 Python 3.9+ 标准库、无需显卡，可 --dry-run 离线试跑。适用于 Claude Code、Codex、OpenClaw、Cursor、WorkBuddy。Talking-head AI voiceover, voice cloning, SRT subtitles.

字幕生成、视频自动加字幕、音频转字幕、音频转文字技能：把本地视频或音频（口播、访谈、课程、Vlog、会议录音、播客）上传识别，生成带时间轴的 SRT 字幕，或导出 TXT 文字稿、JSON（含文本、语言、时间分段）；再用 ffmpeg 或剪映把 SRT 加到视频上。还能在给文字稿配音时同步生成 SRT 和逐句时间轴。提交前自动读取 capabilities 的每秒费率、文件大小和时长上限，费率变化会被服务端拒绝，避免意外扣费。由 Miragea 开放 API 识别，本地只需 Python 3.9+ 标准库、无需显卡。适用于 Claude Code、Codex、OpenClaw、Cursor、WorkBuddy。AI subtitle generator, speech to text, audio to SRT, auto captions.