An open-source agentic video production system that turns an AI coding assistant into a pipeline-driven video studio for research, scripting, asset generation, editing and rendering.
- Estrelas
- 23,782
- Forks
- 2,823
- Issues abertas
- 52
- Último commit
- há 7 horas
O que é?
Hypit gives AI coding agents such as Claude Code and Codex a language and runtime for producing video. Given a reference video, an agent breaks it down into a complete workflow covering footage, word-anchored captions, B-roll and effects, then recreates it with new hosts, scripts or products. Templates or a plain-text description can also serve as the starting point.
The project is built around SVML, a markup language that anchors timing to words rather than seconds. A workflow can be compiled into a finished video using captions, motion graphics and code-rendered visuals alone, so no generation model is required. Model services for image, video and speech are optional and pluggable, and the project documents connections to several providers.
The repository is a TypeScript monorepo with a Node.js 22.15+ requirement, installed as an agent skill via npx. Examples in the repo include UGC ranking clips, podcast clips and street interviews, each with its source file and production notes.
Recursos principais
- 01Clones a reference video into an editable workflow covering footage, captions, B-roll and effects.
- 02Anchors timing to words through the SVML markup language instead of raw seconds.
- 03Generates many variants by reusing a composition and regenerating only the changing parts.
- 04Renders captions, motion graphics and code-based visuals locally without a generation model.
- 05Supports pluggable model services, components and examples.
- 06Installs as a skill for Claude Code, Codex and similar coding agents.
Casos de uso
- Producing many ad or hook variants of the same short-form video for A/B testing.
- Building podcast or interview clips with split-screen layouts and speaker-aware captions.
- Localizing a video into several languages while letting the timing re-flow automatically.
- Creating code-rendered explainer or motion graphics videos without paid generation APIs.
Tópicos
Perguntas frequentes
- O que é hypit?
- A video-as-code toolkit that lets AI agents clone a viral video into an editable workflow and generate variants from it. Hypit gives AI coding agents such as Claude Code and Codex a language and runtime for producing video. Given a reference video, an agent breaks it down into a complete workflow covering footage, word-anchored captions, B-roll and effects, then recreates it with new hosts, scripts or products. Templates or a plain-text description can also serve as the starting point.
- hypit é gratuito? Qual é a licença?
- hypit é open source sob a licença Other. Confira os termos da licença antes de usá-lo em produtos comerciais.
- hypit ainda é mantido?
- Manutenção ativa. O último commit de hypit foi em 19 de out. de 2026.
- Para que posso usar hypit?
- Producing many ad or hook variants of the same short-form video for A/B testing. Building podcast or interview clips with split-screen layouts and speaker-aware captions. Localizing a video into several languages while letting the timing re-flow automatically. Creating code-rendered explainer or motion graphics videos without paid generation APIs.
- Onde consigo hypit?
- O código-fonte e as instruções de instalação estão no GitHub: https://github.com/hypit-ai/hypit
Resumo gerado por IA a partir do README do projeto. As estatísticas do GitHub são atualizadas diariamente.
Projetos relacionados
Open-source platform for turning topics or scripts into AI short videos and AI micro-drama episodes, from storyboard to finished film.
An MIT-licensed Claude agent skill that lets Claude Code, Claude Desktop and Claude.ai create and edit motion-graphics videos with Remotion.
Fair-code workflow automation platform for building AI agents and workflows, combining a visual canvas with custom code and 1500+ integrations.
AutoGPT is an open-source platform for building, running, and scheduling AI agents that carry out multi-step workflows from plain-English descriptions or a visual builder.
Open-source chat interface where an AI agent answers questions with text, native components, or custom interactive UI such as charts, calculators and maps.