跳到主要内容Skip to main content
← Research

NotebookLMNotebookLM

Google
免费+订阅Free + subscription 源锚定Source grounding 音频概览Audio overviews Google生态Google ecosystem 需网络环境Network access required partial lifecycle: active 核验Checked 2026-10-22

NotebookLM 是「源锚定」的研究工作台:你给一组源,它只在这些源里回答+生成音频概览/学习指南; 它不是搜索引擎——是资料集上的问答引擎。

A source-anchored research workbench: you hand it a set of sources and it answers only inside them, plus audio overviews and study guides — it is not a search engine but a Q&A engine over your material.

适合与不适合When to use it, when not to

适合:给定资料集内问答(论文组/尽调/课程)、长材料转音频、源引用可追溯。 不适合:开放网络检索、大陆直连(000)。

Q&A and research organization inside a given source set (paper groups, diligence packs, course material), turning long material into audio overviews, "answers only from your sources" scenarios. Not for: open-web retrieval (it does no web search — the sources are yours), mainland direct access (measured 000).

中文生态China availability 国产 / 国内可直接用 / 备案状态 · 快照domestic / direct access / CAC filing · snapshot 2026-10-22

出品属地Origin
国外出品foreign-made Google(美国)。Google (US).
国内使用Direct access
直连不可达direct blocked notebooklm.google.com 大陆直连实测 000(2026-10-22,代理下 200 可达)——直连不可达,需网络环境。notebooklm.google.com measured 000 direct from mainland (2026-10-22; 200 via proxy) —unreachable without a network path.
备案状态CAC filing
未见于备案名单not in filing list:未见于网信办公示名单(2026-10-22 检索);该产品未面向大陆公众提供服务,备案不适用。not in the CAC public registry (searched 2026-10-22); the product does not serve the mainland public, so filing does not apply.

备案信息以网信办《生成式人工智能服务已备案信息》公示名单为准(按批次滚动更新,本站只抄编号不解读);「大陆可直连」是本站从大陆网络直连实测的结果,不是厂商承诺。GitHub 类站点当日实测有波动,口径写在依据里。Filing status is quoted from the CAC public registry (updated batch by batch; this site copies the number and does not interpret it). "Direct access" is this site's own reachability measurement from a mainland network, not a vendor promise. GitHub-hosted sites fluctuated on the measurement day — the note says so.

固定坐标系Fixed coordinate system

8 个维度dimensions, 与同分区其他对象逐项可比each comparable item by item against others in the same section.

模型与开放条件Model & openness

Gemini 系模型驱动(官方叙事);用户不选模型——产品形态是「源锚定问答」,模型是内部实现。

Driven by Gemini-family models (official narrative); users do not pick models — the product form is "source-anchored Q&A", the model is an internal implementation detail.

运行位置Where it runs

网页(notebooklm.google.com)+ 移动 App;处理在 Google 云端。

Web (notebooklm.google.com) + mobile App; processing on Google's cloud.

本地文件Local files

笔记本与源存账户侧云端;源可上传(PDF/Docs/网页链接/YouTube 等,官方产品面)。

Notebooks and sources stored on the account-side cloud; sources can be uploaded (PDF/Docs/web links/YouTube etc., official product surface).

关机后的任务Tasks after shutdown

问答即时;Audio Overview(播客式音频概览)是生成任务线;无「关机续跑」概念。

Q&A is instant; Audio Overview (podcast-style audio overview) is a generation task line; no "keep-running-after-shutdown" concept.

工具与扩展Tools & extensions

源锚定是产品主轴:回答带源引用、可追问;Studio 线生成音频概览/简报/学习指南等 衍生形态(官方产品面);笔记本内源管理(添加/删除/摘要)。

Source anchoring is the product's main axis: answers carry source citations and support follow-ups; the Studio line generates derived forms such as audio overviews/briefs/study guides (official product surface); in-notebook source management (add/remove/summarize).

上下文与记忆Context & memory

以笔记本(notebook)组织——每个笔记本是一组源+对话;跨笔记本记忆未核验到官方声明。

Organized by notebook — each notebook is a set of sources + conversation; cross-notebook memory not verified to an official statement.

权限与限制Permissions & limits

Google 账号体系;免费档与订阅档的笔记本数/源量上限差异按官方页(未逐条核验,D9:价格不采); 训练用途:官方 FAQ 有「不用你的源训练」类承诺——引用前核 FAQ 原文(未逐条核验)。

Google account system; the notebook/source-count limits differing between free and subscription tiers per the official page (not verified item by item, D9: pricing not collected); training use: the official FAQ carries a "your sources are not used for training"-type promise — check the FAQ original before quoting (not verified item by item).

适合什么任务Best for

适合:给定资料集内的问答与研究整理(论文组/尽调材料/课程资料)、把长材料转成音频概览、 要「答案只从你的源里出」的场景。 不适合:开放网络检索(它不做全网搜索——源是你给的)、大陆直连(实测 000)。

Fits: Q&A and research organization inside a given material set (paper groups/due-diligence materials/course materials), turning long materials into audio overviews, scenarios wanting "answers only from your sources". Does not fit: open-web retrieval (it does no whole-web search — the sources are what you give it), mainland direct connection (measured 000).

头号误解Top misconception

  • NotebookLM 不是搜索引擎——它只在你给的源里回答;「查全网」要用 Perplexity 类产品NotebookLM is not a search engine — it answers only inside the sources you give; for "search the whole web" use a Perplexity-class product
  • Audio Overview 的生成质量随源质量波动——垃圾进垃圾出在源锚定产品里同样成立Audio Overview quality tracks source quality — garbage in, garbage out holds just as hard in a source-anchored product
  • 「不用源做训练」是官方 FAQ 承诺——企业敏感材料仍要按条款与 Workspace 合规面核(未逐条核验)"Your sources are not used for training" is an official FAQ promise — for sensitive enterprise material still check the terms and the Workspace compliance surface (not verified line by line)

未知项清单Known unknowns

  • 各订阅档的笔记本/源量上限(未逐条核验)
  • 训练承诺的条款级落点(FAQ 叙事,未逐条核验)
  • 跨笔记本记忆(未核验到官方声明)
  • 当前驱动模型档位(未逐条核验)
  • Notebook/source-count limits per subscription tier (not verified item by item)
  • Terms-level landing of the training promise (FAQ narrative, not verified item by item)
  • Cross-notebook memory (no official statement verified)
  • The current driving-model tier (not verified item by item)

证据来源Evidence sources

判断可回到以下一手源复核。本站核验日Every judgement can be rechecked against the primary sources below. Verified on 2026-10-22,内容更新日content updated on 2026-10-22。

类型Type名称Name链接Link
officialNotebookLM · 官网(产品与能力叙事)NotebookLM · official site (product and capability narrative)https://notebooklm.google.com
officialNotebookLM · Terms(官方条款,2026-10-22 直抓核验)NotebookLM · Terms (official terms, direct-fetch verified 2026-10-22)https://notebooklm.google/terms
changelogNotebookLM · Release Notes(官方更新线,changelog,2026-10-22 直抓核验)NotebookLM · Release Notes (official update line, changelog, direct-fetch verified 2026-10-22)https://notebooklm.google/release-notes

实测记录Measured runs

本档未跑检索质量评测;结论只覆盖产品能力与来源披露。

实测建议(三条,都指向官方可读页没给的东西):

  • 长 PDF(500+ 页)的源覆盖完整度(会不会静默截断)
  • Audio Overview 对中文源的支持质量
  • 免费档笔记本/源上限的实际值(D9:价格不采;额度实测)

This site has not completed measured runs. The coding-oriented test protocol is at protocols/coding/_protocol.md (the coding task suite does not apply to creation products).

Suggested measurements (three, all pointing at things the official readable pages do not give):

  • Source coverage completeness for long PDFs (500+ pages) (whether it silently truncates)
  • Audio Overview's support quality for Chinese sources
  • Actual values of the free tier's notebook/source limits (D9: pricing not collected; quotas measured)

相关条目Related entries

  • GeminiGemini (chat app)同厂不同产品:Gemini 是开放对话+检索,NotebookLM 是封闭源集问答。Same vendor, different product: Gemini is open dialogue + retrieval, NotebookLM is closed source-set Q&A.
  • ElicitElicit同为文献研究向:Elicit 检索学术库再锚定,NotebookLM 只锚定你上传的源。Also literature-research oriented: Elicit searches academic libraries then anchors, NotebookLM only anchors the sources you upload.
  • 秘塔 AI 搜索Metaso AI Search中文向对照:秘塔学术检索+全网,NotebookLM 源集内深挖。Chinese-leaning counterpart: Metaso does academic retrieval + whole web, NotebookLM digs inside the source set.

本页由This page is generated from the ai-handbook 数据层生成(CC BY 4.0)。data layer (CC BY 4.0). 方法论与坐标系定义见仓库内For the methodology and coordinate definitions, see METHODOLOGY.md.