.llmn-home .reveal { opacity: 1 !important; transform: none !important; transition: none !important; } .llmn-home .demo-inner-desktop, .llmn-home .final-cta-preview-viewport { height: auto !important; overflow: visible !important; } .llmn-home .demo-frame { width: 100% !important; transform: none !important; }

LLM-notebook, your AI notebook can Collect it all…

Any material type; interactive materials with sourced answers; native Agent support that reshapes note-taking.

Get started

LLM-notebook

Multi-type collection

Any material, one-click saving.

Files

PDF, Word, Excel, PPT, Markdown, images and more — drag in and it’s parsed automatically.

URL

Paste a link — the page content is fetched and parsed.

LLM chats

With the extension installed, a “Save chat” button appears — one click saves the conversation.

Requires the browser extension

Agent sessions

Just tell your Agent “Save this to LLM-notebook” — it’s filed automatically.

Requires the skill

Extension / skill installation

LLM chats · extension

Install the browser extension

— chats become one-click savable.

Agent sessions · skill

Paste that line into your Agent — it installs the skill itself. Then just say “Save this to LLM-notebook”.

Install the LLM-notebook skill: https://llm-notebook.com/api/skills/skill.md

Ask your files

Cited answers, @-mentioned files, ask by selecting text — material stops being static and becomes interactive.

🎯

Model answers cite their sources

Every claim traces back to the source text — tap [N] to jump back to the source and highlight it.

📎

@file references

Type @ to pull a file into the conversation context.

✍️

Select text and ask

Select text in the workspace to bring it into the conversation and ask about it.

AI Assistant

What does the [N] after an answer mean?

ThinkFirst, what that [N] is: the in-text citation a model answer marks.,Then what clicking it does: jump back to the source paragraph and highlight it.

First, what that [N] is: the in-text citation a model answer marks.Then what clicking it does: jump back to the source paragraph and highlight it.

That’s the in-text citation — it marks where the conclusion comes from, and clicking it jumps back to the paragraph it cites^[1] and highlights it.

Summarize@Multi-type collectionNew note’s contents

ThinkAll four save in one click and are filed automatically — now summing them up as key points.

First, find “Multi-type collection”: four material types — files, URLs, LLM chats, and Agent sessions. All four save in one click and are filed automatically — now summing them up as key points.

Key points from Multi-type collection: Files, URLs, LLM chats, and Agent sessions — saved in one click and filed automatically.

Explain what this term means:interactive

ThinkThen reading the paragraph it sits in, explaining it in context, and citing the source.

Locating “interactive”: material you can question and cite, not static files. Then reading the paragraph it sits in, explaining it in context, and citing the source.

“Interactive” means materials you can question and cite, not static files^[1] — click the number to jump back to the paragraph it cites and highlight it.

Ask your files

Ask Anything

Native Agent interaction

We’re excited to announce that the native Agent is now live. It lives inside your notes — reading, writing, rewriting, and tidying up material, all without switching tools.

🗂

File management, in your own words

One sentence to organize, search, analyze, or generate a new document — the whole MCP flow runs itself.

🔗

Readable by external Agents

Let any Agent read your material library on its own — no context switching.

✏️

Rewrite a doc with one sentence — you stay in charge

Tell the AI what to change — it proposes word-level edits you accept, reject, or refine.

📝

Save a note in one sentence

Talk to the AI and it turns the conversation into an organized note in your library — revisit and keep creating anytime.

🎨

Generate PPTs & mind maps

Turn organized material into a PPT or a mind map, ready to present.

AI Assistant

Help me polish this product intro

ThinkPolishing line by line: cutting the boilerplate opener, dropping filler, adding what’s missing.

Polishing line by line: cutting the boilerplate opener, dropping filler, adding what’s missing.

Per “concise, accurate, complete”, here is 1 suggestion: red marks deletions, green marks additions — click a mark to apply it.

Organize the related files into folders

ThinkCreating a “Launch plan” folder and archiving the launch files into it.

Creating a “Launch plan” folder and archiving the launch files into it.

Plan ready: create a “Launch plan” folder and archive the related files into it.

create_folderDone · 130ms

move_noteDone · 90ms

✅ Archived to “Launch plan”

Search the library for similar materials

ThinkSearching the library for launch-plan material and reading the key points.

Searching the library for launch-plan material and reading the key points.

Found 3 related materials — reading the key points:

search_materials3 hits · 210ms

read_materialDone · 180ms

Matching material · click to open

competitor-launch-recap.md3 hits

marketing-plan.docx2 hits

product-intro.pdf1 hits

Native Agent interaction

Ask Anything

01 · COLLECT

Everything you scroll past, chat about, or write down lands here

Files, web links, LLM chats, Agent sessions — one click to save, then parsed and filed automatically, ready to search and ask about anytime.

Files

PDF, Word, PPT, Markdown, images and more — drag in and it’s parsed and filed automatically.

PDF

Word

Excel

PPT

TXT

Markdown

HTML

Image

JSON

Interactive materials, sourced answers

Cited answers, @-mentioned files, ask by selecting text — material stops being static and becomes interactive.

Model answers cite their sources

Every claim traces back to the source text — tap [N] to jump back to the source and highlight it.

product-intro.pdf

Conversations, web links, files, and images all land in one Material Library — every item can be questioned and cited.

Ask about a file and the answer cites the original paragraph — click [N] to jump back and highlight it.

Which material types does LLM-notebook support?

ThinkLocating the material-types section and checking citation numbers…

The user asks which material types LLM-notebook supports. I found the entry in product-intro.pdf (conversations, web pages, files, images), checked the citation numbers, and will answer with sources.

Conversations, web pages, files, and images — all four save in one click — tap a number to jump back to the source and highlight.

@file references

Type @ to pull a file into the conversation context.

product-intro.pdf is already in this chat — @ another file to compare the two.

Is this the same approach asAsk your filesNew note?

ThinkComparing Ask your files with the current product intro…

The user wants to know whether Ask your files and the current product-intro.pdf take the same approach. I’ll read the key points of Ask your files, cross-check them against the capability list in the product intro, and answer concisely.

Same approach: the product intro defines the four material types and the citation rules, and answers about a file cite the original paragraph — click [N] to jump back and highlight it. One defines the spec, the other follows it.

Select text and ask

Select text in the workspace to bring it into the conversation and ask about it.

product-intro.pdf

Everything you collect — conversations, web links, files, and images — goes into one Material Library, where every item can be questioned and cited.

Ask about a file and the answer cites the original paragraph — click [N] to jump back and highlight it.

Explain what this term means:“Material Library”

ThinkLooking up “Material Library” and locating the source…

The user selected “Material Library” and asked for an explanation. I’ll locate the term’s passage in product-intro.pdf, explain it in context, and attach the citation numbers.

Where everything you save lives. The product intro lists four material types — conversations, web pages, files, and images — and every item can be questioned and cited.

03 · AGENT

Native Agent support, reimagining what a notebook can be

“You’re a grown-up notebook now — you can handle things on your own.”

File management, in your own words

One sentence to organize, search, analyze, or generate a new document — the whole MCP flow runs itself.

AI assistant

Organize this week’s work files and write a weekly report

ThinkSearching this week’s work files, then planning the archiving scheme.

The user wants to organize this week’s work files and generate a weekly report. I’ll search the library for work files, read them, present an archiving plan, and generate the report once it’s executed.

read_materialDone · 680ms

Organize · 10 loose files, archived by meaningSelect none

New “Requirement design” → move in 4 files: requirements doc, review notes, etc.New “Dev & testing” → move in 3 files: integration, acceptance checklist, etc.New “Project meetings” → move in 2 files: schedule, meeting notesNew “Release” → move in 1 file: release plan

create_folderDone · 130ms

move_noteDone · 90ms

create_noteDone · 420ms

Generated file · click to open

work-notes/20260904/weekly_20260904.md

ThinkFiled 10 loose files into 4 folders; now reading this week’s workload.

With the 10 loose files filed into 4 folders by requirement, dev, meeting, and release, I’ll read all of this week’s work files, analyze them, and compile a weekly report.

Done: 4 folders created and 10 loose files filed by requirement, dev, meeting, and release. Analyzed 5 tasks this week — the report is saved under work-notes/20260904.

Ask AnythingDeepSeek

Readable by external Agents

Let any Agent read your material library on its own — no context switching.

Claude Code

Look up how to use the linux rsync command that I noted in LLM-notebook

ThinkSearching the library for rsync notes, then reading the document.

The user wants the linux rsync usage recorded in LLM-notebook. I’ll call search_materials on the library, open the matching document, and report the usage.

search_materials2 hits · 210ms

Matching material · click to open

linux-command-cheatsheet.md2 hits

read_materialDone · 180ms

rsyncships with --progress a flag that shows detailed transfer progress:

rsync -av --progress source_file destination

Ask Claude Code

Install the LLM-notebook skill in your Agent and grant it library access

Install the LLM-notebook skill: https://llm-notebook.com/api/skills/skill.md

Rewrite a doc with one sentence — you stay in charge

Tell the AI what to change — it proposes word-level edits you accept, reject, or refine.

product-intro.md

✏️ Suggestions · 3

LLM-notebook product intro

We’re excited to announce that the native Agent is now live

It helps you read and write notes, rewrite copy, or tidy up material

When you look something up, it searches first, then answers from the source

AI assistant

Remove redundant phrasing

ThinkReviewing the copy line by line: boilerplate, filler, missing details.

I’ll go through the copy paragraph by paragraph: what can go (boilerplate, filler) and which key detail is missing. Then I’ll turn that into word-level suggestions for the user to approve.

submit_edit_proposal3 spots · 420ms

Per “concise, accurate, complete”, here are 3 suggestions: red marks deletions, green marks additions — click a mark to accept or reject.

product-intro.md

Ask AnythingDeepSeek

Save a note in one sentence

Talk to the AI and it turns the conversation into an organized note in your library — revisit and keep creating anytime.

Turn what we just covered into key points for me

Four points: multi-type collection — conversations, web pages, files, and images in one click; answers about a file cite the source; a native Agent that reads and writes your notes; organize and archive in one sentence. Save this as a note?

Save as note

ThinkWriting the key points into a structured note.

I’ll turn the launch-plan key points into a structured note and save it to the library, so it’s easy to revisit and edit later.

create_noteCreated · 180ms

Saved as note · click to open

product-capabilities.md

Ask AnythingDeepSeek

Pricing

New users get a free 14-day trial plus ¥5 of token credit.

Pick monthly, quarterly, or yearly — and unlock reliable capture, AI Q&A, and interactive editing.

Monthly membership

Flexible billing

¥15/mo

Unlimited material import & Agent calls
¥3 token credit included
2 GB storage
24/7 technical support

FAQ

Common questions about features, security, pricing, and support.

What is the relationship between LLM-notebook and Google NotebookLM?
LLM-notebook and Google NotebookLM are two completely independent projects.
Is my data safe?
Your knowledge base is private to you. We take privacy and data security seriously — for details, see our Privacy Policy.
Free or paid? How does billing work?
Yes — there’s a free tier. Paid plans are monthly ¥15, quarterly ¥40 (about ¥13/month), or yearly ¥132 (about ¥11/month). You can check your subscription status in your account.
Extension vs. web app — what’s the difference?
The extension is the capture point: one click saves a conversation from a chat page. The web app is the full workspace (material library + notes + AI assistant) for sourced answers, organizing, and revising.
Found a bug or have an idea?
If you spot a bug or have an idea while using the app, reach out to us.

Start your AI-native notebook

Get started

LLM-notebook

LLM-notebook product intro

Conversations, web links, files, and images all land in one Material Library — every item can be questioned and cited.

Ask about a file and the answer cites the original paragraph — click [N] to jump back and highlight it.

A native Agent reads and writes your notes: rewrite copy, organize files, generate reports.

Answers cite the original paragraph — click [N] to jump back here and highlight.