Description
The hard part of making an explainer video was never understanding the material — it is turning the material into pictures: script, voiceover, timing, motion, subtitles, a full day for five minutes. Reaching for an AI video generator trades that for a different headache, where the footage invents itself, terms are drawn wrong, and no frame is under your control.
anything2explainer takes another route. It is a skill for Claude Code and Codex: give it a topic ("explain vector databases") or an article, pick the length and language, and it produces a black-canvas motion-graphics explainer with TTS voiceover, word-boundary-aligned subtitles, chapter cards, a top HUD and a bottom chapter progress bar. The point is that every frame is drawn in code with Remotion (React + TypeScript) — no stock footage, no generative video model, no frames lifted from anyone else's work — so formulas, charts and terminology stay exact rather than hallucinated.
What it ships is not a CLI but the whole method an AI agent needs to finish the film: a compilable Remotion template, a primitives and lighting library, tooling for voiceover, storyboard, rendering and quantitative QC, written style and motion specs, a multi-agent division-of-labour protocol, and one complete reference film as the quality bar.
The run goes like this: the agent researches the topic with sources, writes the narration, generates the voiceover and a frame-accurate timeline, storyboards every shot, then dispatches parallel build agents that each write one Remotion component per shot, after which QC agents review the rendered frames against the written criteria. A film takes roughly one to three hours of wall clock, most of it agents building shots in parallel, and you are consulted at exactly four checkpoints.
Topic in, film out: a topic or a document plus your chosen length and language yields a 1280×720 30fps H.264 film with voiceover, subtitles, chapter cards and a progress bar.
Every frame in code: visuals are generated with Remotion rather than stock footage or a generative model, so formulas, charts and terms stay exact and nothing is invented.
Voiceover and subtitles built in: integrated TTS (kokoro and others) with word-boundary-aligned subtitles and a frame timeline, where typography, subtitle budgets and TTS switch with the language.
Parallel multi-agent build: after storyboarding, build agents each write one shot component and render in parallel, so dozens of shots come together in tens of minutes.
Quantitative QC: QC agents score the rendered frames against written visual and motion criteria and send failing shots back before delivery.
Full paper trail: the research doc with sources, narration, storyboard, per-shot source code, QC reports and delivery notes are all kept, so the whole pipeline is auditable beyond the finished film.
Chinese or English: switch language in config and re-time one storyboard to each voiceover; a bilingual RAG reference film ships as the quality bar.
anything2explainer takes another route. It is a skill for Claude Code and Codex: give it a topic ("explain vector databases") or an article, pick the length and language, and it produces a black-canvas motion-graphics explainer with TTS voiceover, word-boundary-aligned subtitles, chapter cards, a top HUD and a bottom chapter progress bar. The point is that every frame is drawn in code with Remotion (React + TypeScript) — no stock footage, no generative video model, no frames lifted from anyone else's work — so formulas, charts and terminology stay exact rather than hallucinated.
What it ships is not a CLI but the whole method an AI agent needs to finish the film: a compilable Remotion template, a primitives and lighting library, tooling for voiceover, storyboard, rendering and quantitative QC, written style and motion specs, a multi-agent division-of-labour protocol, and one complete reference film as the quality bar.
The run goes like this: the agent researches the topic with sources, writes the narration, generates the voiceover and a frame-accurate timeline, storyboards every shot, then dispatches parallel build agents that each write one Remotion component per shot, after which QC agents review the rendered frames against the written criteria. A film takes roughly one to three hours of wall clock, most of it agents building shots in parallel, and you are consulted at exactly four checkpoints.
Features
Topic in, film out: a topic or a document plus your chosen length and language yields a 1280×720 30fps H.264 film with voiceover, subtitles, chapter cards and a progress bar.
Every frame in code: visuals are generated with Remotion rather than stock footage or a generative model, so formulas, charts and terms stay exact and nothing is invented.
Voiceover and subtitles built in: integrated TTS (kokoro and others) with word-boundary-aligned subtitles and a frame timeline, where typography, subtitle budgets and TTS switch with the language.
Parallel multi-agent build: after storyboarding, build agents each write one shot component and render in parallel, so dozens of shots come together in tens of minutes.
Quantitative QC: QC agents score the rendered frames against written visual and motion criteria and send failing shots back before delivery.
Full paper trail: the research doc with sources, narration, storyboard, per-shot source code, QC reports and delivery notes are all kept, so the whole pipeline is auditable beyond the finished film.
Chinese or English: switch language in config and re-time one storyboard to each voiceover; a bilingual RAG reference film ships as the quality bar.
