Highlight notes → local Obsidian → LLM Q&A in context, with long-term memory
Users
5
+1 in 30 days
Rating
0.0 / 5
unchanged in 30 days
Ratings
0
0 in 30 days
Store rank
#262K
#219K among extensions
Version
0.5.2
updated 25 Sep 2026
Users
monthly, 1 months
Latest, 19 Oct 2026:5 usersMove the mouse over the chart for any month
Ratings count
monthly, 1 months
Latest, 19 Oct 2026:0 ratingsMove the mouse over the chart for any month
The Chrome Web Store rounds user counts down to the leading digit, so the user line moves in steps. The ratings count is exact and is the better signal of real adoption.
About Markpilot — Web Notes to Obsidian & AI
Markpilot 把「读到 → 划下 → 记住 → 问透」压缩进浏览器侧栏。
【划词笔记】
在任意网页选中文字,一键记下心得。笔记按页面归档,刷新后自动恢复高亮。
【笔记分级:本页 / 本站】
笔记默认跟随单个页面(自动识别 ?id= 等正文参数,丢弃 utm 等跟踪参数,
不再出现"换了个页面参数笔记就串了"的问题)。也可以一键设为「本站」——
在同域名所有页面的侧栏都能看到,问 AI 时同样注入上下文。
整页与整站随时切换,侧栏分组展示一目了然。
【AI 就地问答】
选中读不懂的段落点「问 AI」,侧栏自动带着选中原文、页面正文和你的笔记向 LLM
提问——模型回答的是"你正在读的这一页",而不是泛泛而谈。支持流式输出、
多轮会话、会话历史管理。
· 提问模板:输入框敲 / 一键总结页面、解释内容、翻译、提取要点
· 推理模型(如 deepseek-flash)的思考过程以可折叠块展示
· 可选辅助模型:记忆提取与翻译可跑在更便宜的第二个模型上,问答仍走主模型
【本地 Obsidian 落盘】
一次授权 vault 目录,之后一键把整页笔记 + AI 问答导出为带 frontmatter 的
Markdown。产物是标准 Markdown 文件,也可以导出到任何 Markdown 笔记目录。
长期记忆自动写入 Markpilot-Memory/,跨会话提问时按需注入。
【备份与迁移】
设置页一键导出全部笔记为 JSON;换浏览器、换设备(或从开发版切换到商店版)
时导入即恢复,数据主权完全在你手里。
【你的 Key,你的数据】
· 自带 API Key:支持主流在线 LLM 服务商、OpenAI 兼容端点,以及完全本地运行的模型
· API Key 只存本机浏览器,没有账号体系,没有任何分析统计
· 页面内容仅在你主动提问时发送给你自己选择的服务商
· 使用本地模型时,数据完全不离开你的设备
【开源】
源码与隐私政策:https://github.com/Atituiset/web-notes-ext
【VS Code 姊妹版】
「Markpilot — Code Notes to Obsidian & AI」把同一套笔记与长期记忆带进编辑器,
记忆经 Obsidian vault 在浏览器与 IDE 之间共享。
界面语言跟随浏览器:简体中文、English、繁體中文、日本語、한국어、
Deutsch、Español、Français、Português (BR)、Русский。
Markpilot compresses "read → highlight → remember → understand" into your
browser's side panel.
HIGHLIGHT & NOTE ANYWHERE
Select text on any webpage and jot down your thoughts in one click. Notes are
filed per page and their highlights are restored automatically on reload.
PAGE-LEVEL AND SITE-LEVEL NOTES
Notes stick to a single page by default — content-bearing query params like
?id= are respected, tracking params like utm_* are dropped, so notes never
bleed across article variants. Flip any note to "site-wide" and it appears in
the side panel on every page of that domain, and feeds the AI context too.
ASK AI IN CONTEXT
Stuck on a paragraph? Select it and hit "Ask AI". The side panel sends the
selected text, the page content, and your own notes to the LLM — so the model
answers about the page you are actually reading, not generic chatter.
Streaming replies, multi-turn threads, and thread history included.
· Prompt templates: type "/" for one-tap summaries, explanations,
translations and key-point extraction
· Reasoning models (e.g. deepseek-flash) show their thinking in a
collapsible block
· Optional auxiliary model: memory extraction and translation can run on a
cheaper second model while chat stays on your main one
LOCAL OBSIDIAN EXPORT
Authorize your vault once, then export a page's notes and AI Q&A as Markdown
with frontmatter in one click. The output is plain Markdown, so any
Markdown-based notes folder works too. Long-term memories are written to
Markpilot-Memory/ and injected into future conversations when relevant.
BACKUP & MIGRATION
Export all notes as JSON from the settings page in one click; import to
restore when switching browsers, devices, or moving from a dev build to the
store build. Your data sovereignty, end to end.
YOUR KEYS, YOUR DATA
· Bring your own API key: works with major hosted LLM providers,
OpenAI-compatible endpoints, and fully local models
· API keys live only in your browser. No accounts, no analytics, no tracking
· Page content is sent only when you explicitly ask, and only to the provider
you chose
· With a local model, nothing ever leaves your machine
OPEN SOURCE
Source code & privacy policy: https://github.com/Atituiset/web-notes-ext
ALSO FOR VS CODE
The sibling extension "Markpilot — Code Notes to Obsidian & AI" brings the
same notes and long-term memory to your editor. Memory is shared between
browser and IDE through your Obsidian vault.
The UI speaks your language: English, 简体中文, 繁體中文, 日本語, 한국어,
Deutsch, Español, Français, Português (BR) and Русский — following your
browser language automatically.