npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@lumageo/luma-cli

v0.0.19

Published

Agent-native CLI for Luma media creation workflows

Readme

luma-cli

Go Version

中文 | English

luma-cli 是给 AI Agent 使用的视频运营与内容制作能力包。它把 PikGeo / Luma 的视频制作能力包装成稳定的命令行原子能力,让 Agent 可以完成爆款仿写、内容研究、脚本改写、语音合成、声音克隆、数字人口播、画中画素材穿插、字幕、BGM、封面和项目产物管理。

它不是给用户手工点来点去的工具,而是给 Agent 配上的一套专业视频制作技能。复杂 workflow 放在 skills/ 说明书里,CLI 本身保持为可组合、可测试、可复用的原子能力。

适合做什么

  • 给 Agent 增加短视频制作能力。
  • 从对标内容研究到脚本仿写,再到数字人口播成片。
  • 管理音色、数字人、素材、字幕、BGM、封面等中间产物。
  • 让运营 SOP 沉淀成 skills/,由 Agent 按步骤执行。

安装

npm install -g @lumageo/luma-cli

首次使用需要登录:

luma-cli auth login <phone_or_account>
luma-cli auth complete
luma-cli auth status

如果你已经有 API key,或老用户还在使用 card key,也可以直接保存:

luma-cli auth login --key <API_KEY_OR_LEGACY_CARD_KEY>

安装本地视频运行时:

luma-cli runtime install ffmpeg

同步 Agent skills:

luma-cli skills sync

等价于:

npx -y skills add zl007700/luma-cli -g -y

后续更新 CLI 和 skills:

luma-cli update

Agent 工具发现

luma-cli tools list
luma-cli tools describe tts.synthesize
luma-cli --json tools describe pip.plan

常用能力:

| 能力 | 命令 | | --- | --- | | 内容研究 | research run, research export, research keywords | | 爆款改写 | script rewrite | | 声音克隆 | voice clone | | 云端资产 | asset list voice, asset list roles, asset upload | | 语音识别 | asr | | 语音对齐 | align | | 语音合成 | tts | | 数字人口播 | lipsync | | 本地素材描述 | material describe, material group list, material group describe, material search, material merge, material understand | | 画中画 | pip scene, pip match, pip plan, pip render | | 字幕 | subtitle | | BGM | bgm mix | | 封面 | cover frame, cover render | | 图片/视频生成 | image generate, video generate | | 项目管理 | project create/use/info, project artifact list/schema |

爆款仿写流程

内置 skill:

skills/luma-workflow-viral-remix/SKILL.md

标准流程:

luma-cli project create viral-demo
luma-cli project use viral-demo

luma-cli research run --role "AI工具创业者,想找适合口播拆解的爆款选题" --output step0_content_research.json
luma-cli research export --input step0_content_research.json --output step0_content_research.csv
luma-cli research keywords --input step0_content_research.json --output step0_keywords.json --csv step0_keywords.csv

luma-cli script rewrite --input source_script.txt --output step1_rewrite.json
luma-cli tts "<rewritten_text>" --voice 男声3 --output step2_tts.wav
luma-cli asset list roles
luma-cli lipsync --avatar <selected_role_name> --audio step2_tts.wav --output step3_lipsync.mp4

luma-cli material group describe vlm_ai --output step4_materials_enriched.json
luma-cli material search --materials step4_materials_enriched.json --query "<rewritten_text>" --limit 8 --output step4_material_matches.json
luma-cli pip scene --segments step4_segments.json --output step4_scene_units.json
luma-cli pip match --scenes step4_scene_units.json --materials step4_materials_enriched.json --mode auto --output step4_material_matches.json
luma-cli pip plan --segments step4_segments.json --materials step4_materials_enriched.json --output step4_picture_in_picture_plan.json
luma-cli pip render step3_lipsync.mp4 --plan step4_picture_in_picture_plan.json --output step4_picture_in_picture.mp4

luma-cli subtitle step4_picture_in_picture.mp4 --output step5_subtitle.mp4
luma-cli bgm mix step5_subtitle.mp4 --output step6_bgm.mp4
luma-cli cover generate step4_picture_in_picture.mp4 --title "封面标题" --count 12 --output-dir step7_covers

声音克隆

上传一段参考音频,保存为后续 TTS 可使用的音色:

luma-cli voice clone ./sample.wav --name my_voice
luma-cli asset list voice
luma-cli tts "这是一段测试口播" --voice my_voice --output step2_tts.wav

查看可用音色和数字人

luma-cli asset list voice
luma-cli asset list roles

voice list 是旧版兼容入口,新流程统一用 asset list voice 查看音色资产。

标准中间产物

建议 Agent 在多步任务中使用这些文件名:

| 步骤 | 文件 | | --- | --- | | 内容研究 | step0_content_research.json, step0_content_research.csv | | 改写 | step1_rewrite.json | | 音频 | step2_tts.wav | | 数字人 | step3_lipsync.mp4 | | 画中画 | step4_segments.json, step4_materials.json, step4_picture_in_picture_plan.json, step4_picture_in_picture.mp4 | | 字幕 | step5_subtitle.mp4 | | 背景音乐 | step6_bgm.mp4 | | 封面 | step7_covers/cover_*.jpg |

内置 Skills

| Skill | 说明 | | --- | --- | | luma-shared | 通用认证、项目、产物和失败处理规则 | | luma-content-research | 内容研究、关键词表和选题整理 | | luma-workflow-viral-remix | 爆款仿写完整流程 | | luma-digital-human | 数字人、TTS、声音克隆相关流程 | | luma-subtitle | 字幕生成、切分、样式和烧录 | | luma-material | 本地素材库、素材组、素材检索和 PIP 匹配 | | luma-assets | 素材上传、选择和复用 |

本地素材库默认放在 ~/.luma/material-library。可以把常用素材组导入到默认库,之后 Agent 只需要按素材组名称引用:

luma-cli material library path
luma-cli material library import ./material_library/groups/vlm_ai --replace
luma-cli material group list
luma-cli material group describe vlm_ai --output materials.json

Luma 将命令行工具和 Agent skills 分开分发:npm 负责安装 luma-cli,skills 通过 skills installer 单独同步。用户通常只需要执行一次全量同步:

npm install -g @lumageo/luma-cli
luma-cli skills sync

luma-cli skills sync 会安装 Luma Agent skills,并写入本地同步记录;luma-cli update 会先更新 npm 包,再同步 skills。更多说明见 docs/SKILLS.md

项目工作区

luma-cli project create demo-video
luma-cli project use demo-video
luma-cli project info

项目目录:

source/     原始素材
audio/      配音和音频
subtitles/  字幕文件
effects/    特效文件
output/     输出视频和封面
tmp/        临时文件

project.json 会记录 history 和 artifacts,方便 Agent 恢复上下文、定位中间产物和复跑步骤。

项目结构

luma-cli 按“命令壳 + 原子能力模块 + skills 说明书”的方式组织:

cmd/luma-cli/             CLI 入口
internal/commands/        命令参数解析、输出和轻量调度
internal/commands/pip_*   画中画 scene/match/render 相关实现
internal/commands/material_* 本地素材库、素材组、素材检索
internal/clientruntime/   ffmpeg、字体、BGM、模板等本地运行时缓存
cloud/                    PikGeo / Luma 后端 API 客户端
project/                  项目目录、history、artifacts
shortcuts/                Agent 可发现的工具说明
skills/                   给 Agent 使用的流程说明书
scripts/                  npm 安装和运行入口

设计原则:

  • CLI 命令保持原子化,不把完整业务流程写死在 Go 代码里。
  • 多步骤视频制作流程放在 skills/,由 Agent 按中间产物逐步执行。
  • 素材理解、语义匹配等高级能力由 Luma 云端服务提供。
  • 本地只保留跨平台稳定的能力,例如文件组织、资源缓存、ffmpeg 渲染、结果整理。
  • 新增能力时优先补 tools describe / shortcuts,让 Agent 能自动发现和理解参数。