infinite-creation skill
infinite-creation takes a novel or script through ComfyUI and ffmpeg to a finished video with voice and subtitles.
- 327 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The project runs a local server, ComfyUI and ffmpeg, writes many files to disk and loads the graphics card.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/infinite-creationDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The project runs a local server, ComfyUI and ffmpeg, writes many files to disk and loads the graphics card.
This is third-party code. Review the repository files before installing.
What it does
infinite-creation is a local studio that turns a novel, script or idea into video without manual editing. The pipeline covers a character and style bible, assets, a storyboard, video per shot, music, subtitles and an ffmpeg merge. The root `SKILL.md` holds a contract for the agent. It describes three modes: launching the project itself, replaying the pipeline without launching it from the `skills/` folder, and plugging in your own ComfyUI workflows. Generation needs ComfyUI with models, Node.js 22 and ffmpeg. Key stages can be approved by hand and a single shot redone. The project resumes after a stop and tracks spending. Besides the built-in agent, Claude Code, Codex CLI, Gemini CLI and others can drive the pipeline through a local MCP stdio server.
Who it is for. Creators of short dramas and story explainers who own a powerful graphics card and ComfyUI and want to produce video from text without a crew.
Good fit when
- A novel or script must become a series of short videos with consistent characters.
- Generation broke halfway and work must continue from the same shot.
- The storyboard is reviewed and single shots are redone before the final video is assembled.
- ComfyUI workflows are wired to an agent so it calls them from a description.
Not a fit when
- There is no graphics card or ComfyUI with the required models.
- Rights to the source novel or plot belong to someone else.
- A quick clip in minutes is needed: the full pipeline takes a lot of time.
Example request
Install infinite-creation, start the studio and build a short video from this script, pausing at storyboard review.Limitations
Before the first run you need Node.js 22 or newer, pnpm 9, Bun, ffmpeg and a working ComfyUI at `http://127.0.0.1:8188` with models for video, images and voice. The one-click ComfyUI installer in the README exists only for Windows; on macOS models are installed by hand. The author ran the project on a graphics card with 20 GB of memory. The README and site are in Chinese, and the English `README_EN.md` is shorter. The last commit dates from November 2026.
How to disable. Stop the project server and delete the clone folder. Delete project folders, the database and generated files separately; ComfyUI and models stay on the system.
FAQ
How do I start the studio?
Clone the repository, run `pnpm install`, then `pnpm run dev`. The studio opens by default at `http://127.0.0.1:4710`.
Can an agent install it from one request?
Yes. The README has ready text: the agent follows the `install-skill.md` link, installs the skill and launches the project.
Is ComfyUI required?
Yes, images, video and voice are not generated without it. The default expected address is `http://127.0.0.1:8188`.
Which agents can drive the pipeline?
Besides the built-in one, settings let you choose Claude Code, Codex CLI, Gemini CLI, Qwen Code, Cursor Agent and others through a local MCP stdio server.
Can the skill set be extended?
Yes. Put a `SKILL.md` into the `skills/` folder and the project picks it up as a new skill.
Discussion
No questions or reviews yet.
Related
SkillsA self-hosted pipeline turns a text script into an MP4 with voiceover and stock footage, and an agent drives it through MCP.
A video editor with a local MCP server: an agent reviews the cut, edits the timeline and starts exports, ComfyUI generates
The agent runs reviewed ComfyUI workflows from the command line: images, video, music, speech and object cutouts.
video-production-skill takes an agent from script to video: slides, speech-verified narration, subtitles and a cover.
A Comfy Cloud plugin for Claude Code: commands to generate images, video, audio and 3D, connected through a hosted MCP server
Skills, a CLI and an MCP server that let Claude Code or Codex make video from a brief: editing, graphics, voiceover, subtitles
README in short
The project joins text analysis, assets, storyboard, video, music and an ffmpeg merge into one reviewable pipeline. It ships two workspaces: the main one and a mode for stylised videos from templates.
README badge
Original SKILL.md
--- name: infinite-creation description: 本地全链路「小说/剧本 → 漫剧/短视频」创作 Agent 解决方案。三种用法:①启动本项目一键跑;②不启动本项目,Agent 按本规范 + skills/ 技能库复刻全流程(剧本→人物圣经/风格→资产生图→分镜→逐镜视频→合并成片);③接入用户自定义的文生图/图生图/视频/语音 ComfyUI 工作流。当用户要克隆/安装/启动本项目、或用本项目方案生成漫剧/短视频、或接入自定义工作流时使用。 --- # Agent无限创作 · 本地全链路漫剧/短视频创作 Agent 解决方案 ## 一、项目介绍 把小说/剧本**全自动**生成带语音的漫剧/短视频成片,不限时长。本地跑通「ComfyUI(MiniMax H3 视频 + Qwen-Image 2.1 文生图/图生图 + Krea2 + Qwen3-TTS)+ Agent + Skill」的全链路。 **核心流程(也是本方案可复刻的方法论)**: ``` 小说/剧本/想法 → 剧本 + 大纲 → 人物圣经(bible)+ 风格指南(style) ← 一致性锚点 → 人物/场景/道具/语音资产生成 → 分镜脚本(长镜头聚合,不拆碎) → 逐镜视频(MiniMax H3 Ref2VA,参考图 + 参考音 → 带语音视频) → 自审重试 → ffmpeg 合并成片 ``` **核心能力**:无限视频时长、自动分镜、项目隔离、人物一致性(canonical 参考图跨镜复用)、技能系统、自迭代技能工厂(工作流 → 技能)、断点续跑、分镜级审查/重生成、分阶段运行、显存维护、在线预览。 **技术栈**:Node.js(Next.js 自定义 server)+ node:sqlite + ComfyUI REST/WebSocket + ffmpeg;Agent 自研 ReAct 循环;前端 Next.js 15 + React 18。 ## 二、快速启动(启动本项目) ### 前置要求 | 依赖 | 最低版本 | 用途 | |------|----------|------| | Node.js | ≥ 22 | 内置 `node:sqlite` | | pnpm | ≥ 9 | 依赖管理 | | Bun | 最新 | `pnpm install` 时自动构建 Agent 内核 | | ffmpeg | 最新 | 合并成片 | | ComfyUI | — | 生成图片/视频/语音必需(默认 `http://127.0.0.1:8188`) |