Video & Animation

Router for all Remotion skills

脚本・企画・歌詞から動画用の絵コンテを作成・修正する。MV、アニメ、実写、広告、ショート動画のカット割り、秒数、カメラワーク、動作、セリフ・音楽・SEを設計し、カット画像、印刷用HTMLコンテ表、CSV、フレーム基準のタイムライン、任意の簡易アニマティックを生成する。「動画コンテ」「絵コンテ」「MVコンテ」「カット割り」などの依頼に使う。

Develop or adapt a reusable brand and optional character reference package for a video series, including identity rules, visual tests, and a versioned reference index.

Fetches the transcript from a video link (YouTube and other yt-dlp-supported sites) and produces a detailed but compact summary covering every main topic. Use when the user gives a video URL and asks for a summary, analysis, or "tell me what this video says"; also use when the user asks about previously summarized videos without giving a link ("what did that video I sent last week say about X", "what's in my archive about Y") — the archive lives under ~/VideoSummaries.

Production-ready 9:16 vertical video editing engine based on Remotion, design system tokens, and viral storytelling rules.

Edite tutoriais e vídeos com apresentador usando HyperFrames, explicações brancas e azuis, detalhes pixelados, câmera lateral, Robust Video Matting e fundos personalizados. Use para roteiro de edição, montagem, revisão, CTA e tela final com som discreto.

Crea videos verticales (Reels/TikTok) de Carli Code (@carli.code) que explican un concepto de programación o IA con slides animados en HTML (código que se escribe y se resalta, esquemas) sincronizados con el guion de SU VOZ, con su cara solo en algunos momentos. Flujo de 3 pasos - 1) preguntas, 2) borrador segundo a segundo + guion de voz, 3) video completo + clips por escena + copy + guion, guardados en "Documentos/Contenido tech skill". Úsala cuando pida un video explicando un concepto, "video con mi voz", "reel de código", "explica X en video", "video con slides y voz", o diga "usa la skill de videos", aunque no nombre la skill. No es para carruseles sin voz (carrusel-tech) ni para guiones sueltos (guion-tiktok).

Runs this studio's whole video pipeline from a brief or idea to the final MP4 — brand kit (new clients only) → write (hook, story, script) → storyboard (concepts, style frames, shots, PDF deck) → voiceover → render (music, sound, captions, final, QC) — stopping for the user's approval at every gate, resuming any video from wherever it stopped, and reporting status. Use when the user wants a new video made end to end ("make a video for…", "new promo/explainer/reel for…"), wants to continue or check on a video ("continue the Stargate video", "where are we?", "what's next?"), or asks for the status of all videos. Orchestrates the stage skills; never does their work itself.

Create animated Roblox Shorts for kids (9:16, ~25-35 s) starring Lulu, a charismatic and cute girl avatar, entirely in English: original kid-friendly script, Lulu's ElevenLabs voice (Jessica), scene built and animated in Roblox Studio through the official command line (no clicks), recording only the Studio window, and a punchy final edit with word-by-word captions, SFX and music.

Finishes a video in this studio — picks or generates music for the storyboard's mood (lift on the reveal, ducked under the voice), places every storyboard sound cue on its frame from the studio's sound library, adds branded word-by-word captions from the voice timings, mixes to social loudness (-14 LUFS), renders the full-quality MP4 and cover image, and runs an automatic quality check. Free tools by default; paid music only when the user asks or approves. Use when the voice-over is approved and the user wants the final video, music, sound effects, captions, mix, export, or "render it". Stage 5 of the studio pipeline (write → storyboard → voiceover → render); never changes visuals or words.

Builds or updates a client's brand kit for this studio (colours mapped to roles, fonts, logo files, tagline, tone, contact and call-to-action details, verified facts, and an asset library of the brand's own images — people, offerings, testimonials, places, work, proof) from their website and any logos, brand guides or graphics the user provides, and saves it once per client at studio/brands/<slug>/ for every later video to reuse. Use when the user starts videos for a new client or brand, gives a website or logo to "get the brand colours", asks to set up, check, fix or refresh a brand kit, or when another studio stage needs a brand kit that doesn't exist yet.

Plans a short marketing video (a Reel, an ad, an explainer or a product demo) from an idea to a ready-to-make plan - the brief, three hooks, the script in timed scenes, the shot list, the brand voice and a pre-production checklist. Use when someone asks to plan, script or storyboard a Reel, a Short, a TikTok, an ad, an explainer or a demo video, or to turn an idea into a video plan.

Porta de entrada para editar um vídeo gravado no celular e transformá-lo em Reel. Use sempre que a aluna pedir para editar, "edita esse vídeo", "faz um reel", "transforma meu vídeo", "edita meu último vídeo", "edita no estilo X", colar o caminho de um vídeo, ou perguntar como editar. Acha o vídeo sozinha, confere se ele serve, pergunta o estilo UMA vez se faltar, delega para a skill do estilo (estilo-celular, estilo-editorial, estilo-explicativo, estilo-colagem, estilo-anuncio) e entrega o vídeo pronto na pasta "prontos".

Edita uma gravação vertical (pessoa falando para a câmera) em um Reel com cara de "editado no próprio celular" — legenda de uma palavra por vez, prints crus acima da cabeça, títulos pequenos, cortes secos, sem pausas nem repetições, no máximo 90 s. Use quando a skill `editar` delegar para o estilo celular, ou quando a aluna pedir "estilo celular", "estilo cru", "parece que editei no Instagram". Para o visual de revista use estilo-editorial; para notícia/explicação use estilo-explicativo; para anúncio use estilo-anuncio.

Edita uma gravação vertical (pessoa falando para a câmera) em um ANÚNCIO / criativo para Meta Ads — headline nos primeiros 5 s e depois ESQUEMAS animados que mostram o que a pessoa diz (fluxos, antes/depois, não ✕ / sim ✓, telas de app, prints reais com contador), legenda palavra a palavra, zonas seguras dos anúncios da Meta, sem pausas nem repetições. Use quando a skill `editar` delegar para o estilo anúncio, ou quando a aluna pedir "anúncio", "criativo", "ad", "para tráfego", "patrocinado", "para vender". Para Reel orgânico use estilo-celular ou estilo-editorial.

Create or refine essay videos that feel like a live word-editor session, with expressive text edits, browser cutaways, and synchronized keyboard sound. Use for social launches of writing and the Kurate video style, rather than generic talking-head or montage-only edits.

Compress existing opaque videos with the user's standard MP4/H.264 profile while preserving resolution, frame rate and duration. Remove audio by default; preserve it when explicitly requested. Use for compression, not animation editing or export from design apps.

Plan, script, build and deliver any video in Remotion with an AI narrator, word-synced captions, the person's own footage, photos and screenshots, real royalty-free music, and a frame-by-frame check before delivery. Covers personal stories (funny or sincere), product promos, explainers, tutorials, recaps, greetings and announcements, vertical (Reels, TikTok, Shorts, WhatsApp Status) or landscape. Use whenever someone wants a video made, improved or re-cut, or says video-xoshaka, even if they don't mention Remotion, narration, captions or music.

Use when an existing assembled video or cut needs revisions, recutting, captioning, enhancement, review, or export. For making a new video from an idea or raw footage, use creating-videos.

Use when asked to make a short product launch / feature demo video (e.g. a 20-second 1080x1080 clip from real product screens) with a coding agent such as Claude Opus 5.5, using a code-driven Remotion pipeline plus optional open-source design, animation, 3D and export libraries.

Text-to-speech for voiceovers and spoken output. Renders narration for product, feature, and demo videos as per-scene audio files with exact durations, timing offsets, and captions, and speaks text aloud. Supports the built-in macOS/Linux voice, local Kokoro, OpenAI, ElevenLabs, Gemini, Cartesia, and Deepgram. Use when the user wants a voiceover, narration, or audio for a video or screencast, wants something read aloud or announced, wants an audio file (.wav/.mp3/.m4a), wants to compare or choose voices, or wants to set up TTS providers and API keys.

>- Make a short narrated film (an explainer, a product intro, a feature announcement, a walkthrough of code or a design) from a prompt. Every frame is drawn in code on a canvas, with a local AI voice (Kokoro), a score made in code, word-timed cues, burned-in captions, and .srt/.vtt files. The flow: prompt, then a storyboard page with the real voice and music that the user iterates on, then the full 1080p render. Real frames are sampled and checked at every step. Use when the user asks to "make a video", "make an intro/explainer/announcement video", "a narrated walkthrough of this PR/feature/repo", "turn this into a video", or wants to change the storyboard, script, voice, music or look of a film made this way.

Produce narrated, animated videos programmatically (any niche, format or visual style) — script → AI voice (ElevenLabs) → word-level alignment → deterministic animation (Remotion + SVG) → synced subtitles, SFX/music mix → validated MP4 + SRT + thumbnail + upload metadata. Use when the user wants to create, plan, restyle or batch-produce explainer/educational/kids/product/story/short-form videos, design a visual style from references, set up a video pipeline, or fix sync/subtitle/render problems in one.

Use when a user wants a new video made from an idea, research, supplied footage, a screen recording, a reference, or generated assets. Use for end-to-end production, including when raw footage is provided; use editing-videos when they only want an existing cut changed or reviewed.