登录

SurfSense

MODSetterApache-2.0活跃维护 有专家点评

本地部署的 NotebookLM 替代品，AI 智能体能自动检索、改写和编辑你的文档，数据不出内网，适合对隐私敏感的研究场景。

Air-gapped, privacy-focused, open-source NotebookLM alternative. AI agent that researches, transforms, and edits your documents. Join our Discord: https://discord.gg/ejRNvftDp9

v2.3.1发布于 2026-11-03 17:01 UTC16,836star +34 / 7 天1,607fork同步于 2026-11-11 05:24 UTC
本地部署隐私保护文档编辑智能体检索增强生成
对比 GitHub 官网agentagentsainotebooklmnotebooklm-alternative

专家点评

上手难度：硬核更新于 2026-11-11

自托管、可断网的 NotebookLM 替代品，把文档问答和引用溯源留在自己机器上

适用场景

  • 处理不能出内网的合同、病历、内部报告：索引和模型都能留在本地，换用云端 RAG 服务等于把原文交出去
  • 需要带引用溯源的文档问答，而不是让模型自由发挥：它按页/段回引原文，适合做尽调、法务复核这类要能点回出处的活
  • 在 Word 上做批注式改写：它走修订模式、原文件不动，比让模型直接吐一段文本再手工合并省事
  • 把网页、YouTube、Reddit 抓下来和本地文档混在一个知识库里查：内置抓取插件，省掉自己拼爬虫加切分管线

踩坑

  • 想全本地跑模型时容易卡住：issue 里本地 LLM 支持、Ollama 连不上被反复提出，说明本地模型接入不是开箱即用，得预留调试时间
  • 全新 Docker 安装后注册/登录失败是反应数最高的 bug 之一，all-in-one 容器的后端 URL 配置也有单独 issue，第一次部署别指望一条命令就通
  • README 明说的前置依赖容易漏：Office 转 PDF 要另装 LibreOffice（镜像里没有），图像生成默认没配模型，官方插件要邮箱领许可证，这三处任一没配都会表现为功能静默不可用

同类对比

要的是成熟托管体验、不在乎数据出网，直接用 NotebookLM；要纯检索层、不需要文档编辑和抓取，LlamaIndex / RAGFlow 这类更轻，SurfSense 的价值在把抓取、索引、修订编辑打包成一个应用

怎么上手

由 README 提炼，执行前请确认
原文 
curl -fsSL https://raw.githubusercontent.com/MODSetter/SurfSense/main/docker/scripts/install.sh | bash
硬件要求按 README 说明

README 未说明具体显存或 GPU 型号要求；可完全离线运行，模型可选本地或远程，本地模型需自行准备。

支持 Apple Silicon
下载：源码 zip·GitHub Releases（安装包）本站不镜像安装包

项目说明

由 README 自动翻译提炼

本地离线运行的 NotebookLM 替代品，用 AI 代理检索、转换和编辑你的私密文档

本地部署的 NotebookLM 替代品，AI 智能体能自动检索、改写和编辑你的文档，数据不出内网，适合对隐私敏感的研究场景。

核心能力

  • 索引保存在本地磁盘，可选择本地或远程模型
  • 支持 PDF、Word、Excel、PPT 等格式的检索与引用回答
  • 以修订和批注方式编辑 Word，原文件始终不改
  • 内置网页、YouTube、Reddit 等插件抓取公开内容
  • 支持 MCP 第三方插件，可按 URL 自行添加

容易踩的坑

Office 格式转 PDF 需在设置中开启 LibreOffice 支持，安装包不含该组件；图像生成需另选本地或远程图像模型，默认未配置；SurfSense 官方插件需许可证，可在应用内用邮箱领取 14 天免费试用。

内容由模型根据 README 生成，可能不完整或不准确，以官方文档为准。看原文 README

最近发版

5 个版本v2.3.1
全部 release
  1. v2.3.1最新正式版2026-11-03 · 2026-11-03 17:01 UTC

    Known issues

    • Mac users on 2.3.0 or earlier: Your current version can't install this update. Download 2.3.1 from the downloads page and install it. Your data stays as it is. After that, updates install in the app as usual.
    • Windows and Linux update in the app as before.

    What's Changed

    Full Changelog: https://github.com/MODSetter/SurfSense/compare/v2.3.0...v2.3.1

  2. v2.3.02026-11-02 · 2026-11-02 19:32 UTC
    查看更新说明

    Known issues

    • Image edits are slow on a processor. Editing an image locally without a graphics card took 11 to 16 minutes in testing, about twice as long as drawing one.
    • No GPU in Docker yet. The Docker image runs local models on the processor only.
    • A DRM-protected EPUB is not refused at upload. MOBI and AZW3 books with DRM, and KFX books, are refused; an EPUB is not checked for DRM.

    What's Changed

    New Contributors

    Full Changelog: https://github.com/MODSetter/SurfSense/compare/v2.2.5...v2.3.0

  3. v2.2.52026-10-27 · 2026-10-27 01:28 UTC
    查看更新说明

    Known issues

    • Going back to v2.2.4 needs no restore. This release does not change the database.
    • Instagram shows what a logged-out visitor sees. A profile or its reels page lists its latest 12, with no Show more, and a post is added without its comments.

    What's Changed

    Full Changelog: https://github.com/MODSetter/SurfSense/compare/v2.2.4...v2.2.5

  4. v2.2.42026-10-25 · 2026-10-25 23:53 UTC
    查看更新说明

    Known issues

    • The update to v2.2.4 is a full download. Earlier versions download every update whole; updates after this one download only what changed.
    • SSE plugins stop working. opencode 2.x speaks Streamable HTTP only. Remove an SSE plugin and add the provider's /mcp address instead.

    What's Changed

  5. v2.2.32026-10-21 · 2026-10-21 10:21 UTC
    查看更新说明

    Known issues

    • The first start needs free disk space. 2.2.3 upgrades your library database on its first start (migrations 0031 and 0032) and saves a copy of it first. It needs free space about the size of ~/.surfsense/surfsense.db. If there isn't enough, SurfSense closes without a message. Free some space and start it again.
    • Going back to 2.2.2 needs a manual restore. 2.2.2 can't open a database that 2.2.3 has upgraded. Before you reinstall 2.2.2, quit SurfSense and replace ~/.surfsense/surfsense.db with the copy saved at the upgrade, ~/.surfsense/backups/0030-0032.db. If surfsense.db-wal or surfsense.db-shm is still there, delete it too. Anything you added after updating won't be in the copy, and that includes sources, chats and notes. If you updated straight from 2.2.1, the copy is 0029-0032.db.

    What's Changed

    New Contributors

    Full Changelog: https://github.com/MODSetter/SurfSense/compare/v2.2.2...v2.2.3

用户点评

登录用户的真实使用记录，可按版本和硬件筛选
共 0 条

    还没有人点评。第一个说出真实体验的人最有价值。

    README 原文

    未翻译
    在 GitHub 打开

    Stars Latest release License Discord

    Redline a contract with tracked changes and comments

    https://github.com/user-attachments/assets/88326b9d-e763-46db-9635-75f9df9e8682

    Add sources from the web, then turn them into a web page

    https://github.com/user-attachments/assets/073577b7-f055-4f59-8540-330bd764aeb0

    SurfSense is a free, open-source desktop app for documents you can't upload. A privacy-focused, NotebookLM-style AI agent that researches, transforms and edits your documents, on your machine. The index stays on your disk, you pick the model, local or remote, and there is no account.

    PlatformDownload
    Windows x64SurfSense-Setup.exe
    macOS Apple SiliconSurfSense-arm64.dmg
    Linux x64AppImage · deb
    Docker Linux, macOScurl -fsSL https://raw.githubusercontent.com/MODSetter/SurfSense/main/docker/scripts/install.sh | bash
    Docker Windows (Docker Desktop)irm https://raw.githubusercontent.com/MODSetter/SurfSense/main/docker/scripts/install.ps1 | iex

    [!NOTE] Docker self-hosts it on a server you open from any device on your network. Run the same command again to update.

    Quick start

    1. Install SurfSense and do the initial onboarding setup.
    2. Drop your files or folders in your sources.
    3. Ask the agent to search or extract or edit your files.

    What it does with each format

    FormatResearch (cited answers)CreateEdit (as a copy)Transform
    PDF✓ scanned pages too³✓Fill and flatten forms; stamp watermarks, page numbers, headers, footersMerge, split, extract, rotate, reorder pages
    Word .docx✓✓ or from your own .docx as a template✓ body text, as tracked changes and commentsTo PDF¹
    Excel .xlsx✓ the agent also reads formulas✓ with formulas and native charts✓ cell values and formulasAnalyse with pandas, chart the results
    PowerPoint .pptx✓✓ or from your own .pptx as a templateReplace text on slides and notes; delete or duplicate slidesTo PDF¹
    CSV✓——Analyse with pandas, chart the results
    Images PNG, JPEG, TIFF, BMP, WebP✓ text read by OCR³Images and infographics, edited by instruction²—Place in new documents
    Markdown, plain text✓✓ Markdown——
    HTML✓✓ web page——
    Books EPUB, MOBI, AZW3✓ without DRM———
    Recordings WAV, MP3, FLAC, Ogg Vorbis✓ transcribed on your machine———
    Podcast (audio)—✓ with a transcript, voiced on your machine by default——
    Mind maps, flashcards, quizzes—✓ in the app——

    ¹ With Office support (LibreOffice), turned on in Settings; it is never in the installer. ² Needs an image model, local or remote; none is chosen by default. ³ OCR reads Latin, Chinese and Japanese script.

    Edits, conversions and PDF tools always make a new file; your original is never changed. PDF, Word and Excel do the most today, and support for the other formats keeps improving.

    Plugins

    Connect them in Settings › Plugins.

    By SurfSense. They read public pages without your own accounts and need a SurfSense license: download the app and enter your email there for a free 14-day one.

    PluginWhat it does
    Google SearchSearch Google and read its results, ads and related questions
    Web CrawlRead a web page, or crawl a whole site into clean text
    YouTubeRead YouTube videos, channels and their comments
    TikTokRead TikTok videos and their comments, search users, and see what is trending
    RedditRead public Reddit posts with their comment threads, or a whole community
    InstagramRead public Instagram profiles, posts and reels, and find profiles by search
    Google MapsFind places on Google Maps and read their reviews
    AmazonRead Amazon products and search results
    WalmartRead Walmart products, search results and reviews
    IndeedRead job listings on Indeed

    Third-party. MCP servers from the plugin registry, such as DeepWiki and Notion, or any you add by URL. No license needed; they stay off until you allow them.

    How it compares

    NotebookLM (now Gemini Notebook)SurfSense
    Runs offline / air-gappedNoYes
    Your documents leave your machineYesNo, unless you pick a remote model
    Can edit your filesNoYes
    Open sourceNoApache-2.0
    ModelsGemini onlyAny OpenAI-compatible API, or a local one

    Community

    License

    Apache-2.0. See LICENSE.

    相关项目

    同分类里 star 最高的项目
    对比全部 →
    langchain
    langchain-ai·Python

    用 Python 搭建 LLM 应用和智能体的开发框架，把模型调用、工具、检索和记忆串成可组合的链路，省去自己拼装底层胶水代码。

    智能体开发+4RAG 框架进阶
    149338+19625klangchain==1.4.102026-11-10 21:33 UTC
    awesome-llm-apps
    Shubhamsaboo·Python

    收录 100 多个可直接上手的 AI Agent、Agent Skills 和 RAG 应用示例，帮开发者快速找到能跑起来的 LLM 项目参考。

    AI Agent+4RAG 框架进阶
    144387+32021k2026-11-10 01:54 UTC
    hello-agents
    datawhalechina·Python

    从零开始手把手教你构建 LLM 智能体的中文教程，覆盖 Agent 原理、RAG 等核心实践，适合入门与动手复现。

    智能体教程+4RAG 框架进阶
    86648+33911kV1.0.32026-10-13 22:28 UTC
    serviette
    pathwaycom·Python

    一个免代码的 RAG 服务器，能随文档更新自动同步答案，支持 8 种向量数据库，基于 Pathway 构建。

    RAG+4RAG 框架进阶
    587131.5kv0.1.82026-11-10 17:46 UTC