← 返回 Skills 市场
antonia-sz

Notebooklm

作者 antonia huang · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
590
总下载
0
收藏
4
当前安装
1
版本数
在 OpenClaw 中安装
/install notebooklm-py
功能描述
Google NotebookLM 非官方 Python API 的 OpenClaw Skill。支持内容生成(播客、视频、幻灯片、测验、思维导图等)、文档管理和研究自动化。当用户需要使用 NotebookLM 生成音频概述、视频、学习材料或管理知识库时触发。
使用说明 (SKILL.md)

NotebookLM Skill

通过非官方 Python API 访问 Google NotebookLM 的全部功能。

前置要求

pip install notebooklm-py
pip install "notebooklm-py[browser]"  # 首次登录需要
playwright install chromium

认证

首次使用需要登录:

notebooklm login

或使用 Edge(某些企业环境需要):

notebooklm login --browser msedge

核心功能

1. 笔记本管理

# 创建笔记本
notebooklm create "我的研究"

# 列出所有笔记本
notebooklm list

# 切换当前笔记本
notebooklm use \x3Cnotebook_id>

# 删除笔记本
notebooklm delete \x3Cnotebook_id>

2. 添加来源

# 添加网页
notebooklm source add "https://example.com/article"

# 添加本地文件(PDF、Word、Markdown、音频、视频、图片)
notebooklm source add "./paper.pdf"
notebooklm source add "./lecture.mp3"

# 添加 YouTube 视频
notebooklm source add "https://youtube.com/watch?v=xxx"

# 执行网络研究并自动导入
notebooklm source add-research "人工智能发展趋势" --mode deep

3. 内容生成

# 生成音频概述(播客)
notebooklm generate audio "让内容更生动有趣" --wait

# 生成视频概述
notebooklm generate video --style whiteboard --wait

# 生成电影风格视频
notebooklm generate cinematic-video "纪录片风格总结" --wait

# 生成幻灯片
notebooklm generate slide-deck

# 生成信息图
notebooklm generate infographic --orientation landscape

# 生成测验
notebooklm generate quiz --difficulty hard --quantity 10

# 生成记忆卡片
notebooklm generate flashcards --quantity 20

# 生成思维导图
notebooklm generate mind-map

# 生成数据表格
notebooklm generate data-table "对比主要观点"

# 生成报告
notebooklm generate report "研究简报"

音频格式选项

  • deep-dive - 深入探讨
  • brief - 简要概述
  • critique - 批判性分析
  • debate - 辩论形式

视频风格选项

  • whiteboard - 白板风格
  • news - 新闻风格
  • cinematic - 电影风格

4. 下载生成的内容

# 下载音频(播客)
notebooklm download audio ./podcast.mp3

# 下载视频
notebooklm download video ./overview.mp4

# 下载幻灯片(支持 PDF 和 PPTX)
notebooklm download slide-deck ./slides.pdf
notebooklm download slide-deck --format pptx ./slides.pptx

# 下载测验(支持 JSON、Markdown、HTML)
notebooklm download quiz --format json ./quiz.json
notebooklm download quiz --format markdown ./quiz.md

# 下载记忆卡片
notebooklm download flashcards --format json ./flashcards.json

# 下载思维导图
notebooklm download mind-map ./mindmap.json

# 下载信息图
notebooklm download infographic ./infographic.png

# 下载数据表格
notebooklm download data-table ./data.csv

5. 聊天问答

# 向笔记本提问
notebooklm ask "总结核心观点"
notebooklm ask "解释这个概念"

# 使用自定义人格
notebooklm ask "用简单的语言解释" --persona "友好导师"

完整工作流示例

学术研究转播客

# 1. 创建笔记本
notebooklm create "论文解读"

# 2. 上传论文 PDF
notebooklm source add "./research-paper.pdf"

# 3. 生成播客(深入探讨风格)
notebooklm generate audio "深入分析研究方法" --format deep-dive --wait

# 4. 下载 MP3
notebooklm download audio ./paper-podcast.mp3

批量视频生成

# 1. 创建笔记本并添加多个来源
notebooklm create "课程笔记"
notebooklm source add "./lecture1.pdf"
notebooklm source add "./lecture2.pdf"
notebooklm source add "https://reference-site.com"

# 2. 生成白板风格教学视频
notebooklm generate video --style whiteboard --wait

# 3. 同时生成测验检验学习效果
notebooklm generate quiz --difficulty medium

# 4. 下载所有材料
notebooklm download video ./lesson.mp4
notebooklm download quiz --format markdown ./quiz.md

研究项目自动化

# 1. 执行深度网络研究
notebooklm create "市场调研"
notebooklm source add-research "2024年电动汽车市场趋势" --mode deep

# 2. 生成综合分析报告
notebooklm generate report "市场分析报告"

# 3. 生成数据对比表格
notebooklm generate data-table "对比主要厂商的市场份额"

# 4. 生成演示幻灯片
notebooklm generate slide-deck

# 5. 批量下载所有材料
notebooklm download report ./report.md
notebooklm download data-table ./data.csv
notebooklm download slide-deck ./presentation.pdf

Python API 使用

当需要更复杂的逻辑时,使用 Python API:

import asyncio
from notebooklm import NotebookLMClient

async def research_workflow():
    async with await NotebookLMClient.from_storage() as client:
        # 创建笔记本
        nb = await client.notebooks.create("自动化研究")
        
        # 批量添加来源
        await client.sources.add_url(nb.id, "https://example.com/1", wait=True)
        await client.sources.add_file(nb.id, "./doc.pdf", wait=True)
        
        # 生成内容
        status = await client.artifacts.generate_audio(
            nb.id, 
            instructions="制作引人入胜的播客",
            format="deep-dive"
        )
        await client.artifacts.wait_for_completion(nb.id, status.task_id)
        
        # 下载
        await client.artifacts.download_audio(nb.id, "output.mp3")
        
        return nb.id

# 运行
nb_id = asyncio.run(research_workflow())
print(f"完成!笔记本ID: {nb_id}")

故障排查

# 检查认证状态
notebooklm auth check --test

# 查看元数据
notebooklm metadata --json

# 查看分享状态
notebooklm share status

# 列出支持的语言
notebooklm language list

注意事项

  • 这是非官方库,使用 Google 内部 API,可能随时变动
  • 大量使用时可能触发 rate limit
  • 适合原型、研究和个人项目
  • 首次登录后,凭据会保存在本地,后续无需重复登录
安全使用建议
This skill appears internally consistent with an unofficial NotebookLM CLI/client. Before installing, be aware: (1) notebooklm-py is unofficial and uses browser-based login; credentials are stored locally by that library—only proceed if you trust the package and workflow. (2) Playwright will download a browser runtime (Chromium). (3) The tool can upload local files and perform web research on your behalf—do not add sensitive documents you don't want sent to Google/internal services. (4) The included scripts simply call the notebooklm CLI via subprocess; they don't add hidden network endpoints. If you need stronger assurance, inspect the notebooklm-py package source (on PyPI/GitHub) and verify its authentication/storage behavior and network endpoints before use.
能力评估
Purpose & Capability
Name/description match the behavior: SKILL.md and scripts revolve around the notebooklm CLI and Python client (creating notebooks, adding sources, generating and downloading artifacts). No unrelated binaries or credentials are requested.
Instruction Scope
Instructions stay within the stated scope: installing the unofficial package, running browser-based login, adding web/local sources, generating content, and downloading artifacts. The scripts call the notebooklm CLI via subprocess, which is consistent with the described workflows. There are no instructions to read unrelated system files or exfiltrate data outside the NotebookLM tool.
Install Mechanism
No registry install spec is present; SKILL.md recommends pip installing notebooklm-py and Playwright (and running 'playwright install chromium'). These are normal, traceable package operations and the included code files don't download arbitrary external archives or use obscure URLs.
Credentials
The skill declares no required environment variables or credentials. It requires interactive browser login (the library stores credentials locally per its notes), which is appropriate for accessing NotebookLM. There are no unexplained secret requests.
Persistence & Privilege
always is false and the skill does not attempt to modify other skills or global agent settings. Autonomous invocation is allowed (platform default) but not excessive for this skill's purpose.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install notebooklm-py
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /notebooklm-py 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Initial release of the notebooklm-py skill providing access to Google NotebookLM via an unofficial Python API. - Supports notebook management (create, list, switch, delete) - Enables adding diverse content sources: web links, local files (PDF, audio, video), YouTube, and automated research - Offers one-command generation of materials: podcast/audio, video (whiteboard, cinematic), slides, infographics, quizzes, flashcards, mind maps, data tables, and reports - Allows content download in multiple formats (mp3, mp4, pdf, pptx, json, md, csv, png) - Provides chat/question-answering with custom personas - Includes full workflow examples and Python API usage - Documentation covers installation, authentication, troubleshooting, and usage notes
元数据
Slug notebooklm-py
版本 1.0.0
许可证 MIT-0
累计安装 4
当前安装数 4
历史版本数 1
常见问题

Notebooklm 是什么?

Google NotebookLM 非官方 Python API 的 OpenClaw Skill。支持内容生成(播客、视频、幻灯片、测验、思维导图等)、文档管理和研究自动化。当用户需要使用 NotebookLM 生成音频概述、视频、学习材料或管理知识库时触发。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 590 次。

如何安装 Notebooklm?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install notebooklm-py」即可一键安装,无需额外配置。

Notebooklm 是免费的吗?

是的,Notebooklm 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Notebooklm 支持哪些平台?

Notebooklm 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Notebooklm?

由 antonia huang(@antonia-sz)开发并维护,当前版本 v1.0.0。

💬 留言讨论