html-video CLAUDE.md

html-video CLAUDE.md is an instructions file for Claude Code from nexu-io/html-video. It costs 6,229 tokens per session, scanned A, original, Apache-2.0.

Project instructions for html-video, an open-source tool that lets a coding agent create HTML-based videos through several rendering engines. A rendering engine is the software that turns the project files into video frames.

In plain words
What is it for?
Use them when designing the engine adapters, agent skills, templates, documentation, or product structure for HTML video creation.
Why use it?
They clarify the project's role, its boundaries, and how it differs from using one video engine directly.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md. Also seen: mentions CLAUDE.md.

About the project

html-video is a local tool that turns HTML, CSS, and data into animated MP4 videos using configurable rendering engines and templates. Coding agents use it to create and export videos on a user's machine, with optional AI-generated soundtracks. The catalogue skills operate the tool and its video-generation workflow.

nexu-io/html-video · 4,525 stars · on GitHub · open-design.ai

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/nexu-io/html-video/claude-md
Clone the repo
git clone --depth 1 https://github.com/nexu-io/html-video

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for html-video CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/nexu-io/html-video/claude-md.svg)](https://agentmods.dev/instructions/nexu-io/html-video/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/nexu-io/html-video/claude-md"><img src="https://agentmods.dev/badge/instructions/nexu-io/html-video/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 6,229 This file is loaded in full into every session.
When invoked 6,229 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.06229 $0.06229
Opus 5 $0.03114 $0.03114
Sonnet 5 $0.01246 $0.01246
Haiku 4.5 $0.00623 $0.00623

Measured 6d ago against content hash 1cd732aaae77, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

html-video CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 224 lines

How it starts

The opening of the file, as written. The whole thing — 224 lines — stays where its author put it; the contents beside it link to each section on GitHub.

html-video 项目工作区

Open-source HTML→Video meta-layer。让本地 coding agent 跨多个渲染 engine(Hyperframes / Remotion / Motion Canvas / Revideo)一站式做 HTML 视频。 启动时间:2026-05-26(继 OD / HA / OD-pitch / OD-landing 之后的下一代 nexu-io 开源产品)

角色与边界

  • 角色:T9(端口段 3071-3079)—— 现有 T1-T8 已分配
  • 职责:html-video 项目的产品定位 / 架构 / engine 适配器 spec / 模板生态 / agent skill 设计 / 文档 / 公开 launch 物料
  • 不做
    • 跨项目协调(T1 主控的活)
    • OD 主线开发(T5 open-design 工作树)
    • 增长策略 / OKR(T2 opendesign-growth

产品定位(决策时间线)

  • 2026-05-26 启动思路:Joey 最初以为 Hyperframes 是闭源 SaaS,想做 "HTML 视频开源整合工具"。调研发现 Hyperframes 已经是 Apache-2.0 开源 + 21K★ + agent-native,跟 Joey 设想的产品 1:1 重合。
  • 2026-05-26 定位拍板:放弃"做 HF 杀手"的正面竞争路线,改走 Meta-aggregator —— 把 HF / Remotion / Motion Canvas / Revideo 都包成可选 backend,agent 选 engine + 模板 + 一键出片。差异化明确,跟 HF 不正面撞。

关键差异化 vs Hyperframes

维度 Hyperframes html-video
渲染引擎 单一(GSAP+Puppeteer) 多引擎 pluggable
Authoring 范式 单一(HTML+CSS+GSAP) 跟随 backend(HTML/React/TS-generator 都行)
Agent 能力 自家 skill 跨引擎 agent 决策 + skill 注入
模板生态 自家曲库 跨生态 curated + 社区贡献

目录结构

html-video/
├── README.md                  公开门面(对外,已立)
├── CLAUDE.md                  本文件,内部 working notes
├── LICENSE                    Apache-2.0
├── .gitignore                 通用
├── notes/                     迭代笔记 / RFC 草稿 / 决策记录
├── research/                  竞品调研 / engine spec 草案 / API 探索
└── assets/                    素材(截图 / 草图 / launch 物料)

代码骨架(adapter spec、CLI、studio 等)等架构定型后另起 core/ adapters/ cli/ studio/ 等目录。

当前状态(2026-05-28 v0.8 content-graph + multi-frame)

v0.7 之前的进展见 git log。本节记 v0.8 落地的事 + 还没接的事。

  • @html-video/content-graph 新 package:ContentGraph schema(entity/data/text 节点 + sequence/dependency/contrast 边)+ validate + topoSort(dependency 硬约束 / sequence 软排 / contrast 不参与排序)+ totalDurationSec
  • ✅ core:ProjectcontentGraphPath + frames[]orchestratorwriteContentGraph / readContentGraph / writeFrameHtmlexportMp4 多帧路径走 ffmpeg concat(缺 ffmpeg 时报友好错)
  • ✅ studio agent:prompt 同时教单帧 fast path + 多帧(```json#content-graph + ```html#<nodeId> 双块协议);server 解析后自动调 writeContentGraph + writeFrameHtml;无 graph 时回落 v0.7
  • ✅ studio UI:frames-strip 切帧 + graph viewer modal(JSON 只读 + 下载);单帧 fast path 时整条隐藏
  • ✅ 已 push nexu-io/html-video main(commit 149ace9
  • ⏳ HF upstream 真实 render(v0.9?):adapter-hyperframes 仍是 stub,单帧/多帧 export MP4 都是空文件
  • ⏳ ffmpeg concat 真实跑通:依赖 ffmpeg 装好;adapter render 出真 MP4 之后才 end-to-end 有用
  • ⏳ RFC-06 正式文档稿(落 research/2026-05-28-spec-06-content-graph.md);目前规范散在 content-graph package README + smoke + agent prompt 三处
  • ⏳ studio 的 graph viewer 现在只读,未来可加可视化编辑(拖节点 / 加边)

Read the full file on GitHub · 224 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 6d ago First seen · 224 lines · 6,229 tokens per session scan A 1cd732aaae77

Subscribe to this mod's changes

html-video CLAUDE.md is an instructions file published in the GitHub repository nexu-io/html-video (4,525 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 6,229 tokens to every session, about $0.0311 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.