LLM-notebook: the notebook for the AI era

Notes need AI, and AI needs notes

In the AI era information is never scarce — what is scarce is information you can understand, find and reuse. Notes and AI are two sides of the same thing: AI brings your notes to life, and your notes make AI understand you. The more you use it, the better it knows you: a loop that feeds itself.

Notes need AI. Filing by hand is expensive — manual sorting, manual summarizing, lots saved and little reused. Knowledge you bothered to keep should not be single-use. AI lets notes identify themselves, generate structure and distill the key points, and be called on again later. You do the recording and the judging; AI does the organizing, connecting and retrieving.

AI needs notes. A general model has public knowledge, but not your experience, your judgment or the questions you actually care about. Notes carry that private context. Without them, AI can only give you an average answer; with them, it can give you your answer. Notes are AI's memory layer; AI is the amplifier for the value of your notes.

That thinking is why we built LLM-notebook — a notebook for the AI era. More than a place to record: it builds a context for your thinking that can be understood, talked to and reused. Every note becomes the starting point for the next thought.

Four kinds of material, one place

Material never comes in just one shape. LLM-notebook opens the intake wide — a dozen common file formats, web links, LLM chats and Agent sessions all go straight in:

Type	How to import
File	"New material" → "File". Upload PDF / Word / PPT / Markdown / images / Excel / audio / archives and other common formats
Web links	"New material" → "Web links". Paste a URL and the page is fetched at high fidelity
LLM chats	Via our browser extension (, ). Click the button on the chat page and the whole conversation is saved
Agent session	Via our skill. The first use asks you to grant the Agent read and write access to LLM-notebook

LLM chats and Agent sessions are the two kinds of material that only exist in the AI era. What they record — the follow-up questions, the corrections, the conclusions — is knowledge from working with a model, and it is reusable in its own right. It used to live only in a chat log and vanish the moment you closed the page. Now it is filed, searchable and citable like any file.

• Install the extension:
• Install the skill: send this line to your Agent and it installs itself

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

Material you can question, answers you can trace

With LLM-notebook's native AI, material stops being a static, read-only object and becomes something you can question and discuss. Answers are grounded in the material: that adds your personal context on top of the model's general knowledge, and it also makes every answer traceable, which cuts hallucination sharply.

• Answers with citations: claims point back to the paragraph; click [N] to jump there, highlighted.
• @ a file: type @ in the input box to pull that item into the current context.
• Ask from a selection: select a passage in the workspace and ask about it directly.

The notebook gets things done: native Agent integrated

LLM-notebook ships with a native Agent, which moves the notebook from "a container for content" to "something that gets tasks done":

• Edit a document in one sentence: word-level suggestions, red for deletions and green for additions. Accept, reject or undo block by block.
• File management: one sentence to organize, search, analyze or generate a new document — "organize this week's work files and draft a weekly report", for example. Archiving plans are confirmed before anything runs.
• Long-term memory: it remembers your writing style, your working languages and your background, so answers fit your habits.
• Generate PPT / mind maps: organized material becomes editable artifacts instead of read-only previews.
• Readable by external Agents: an Agent with the LLM-notebook skill installed can read and act on your library — an external brain.
• Native MCP capabilities: searching your whole library, searching the web, and more to come.

Compared with Gemini Notebook (formerly NotebookLM)

LLM-notebook and NotebookLM (Google) are two independent projects, despite the similar names (purely a coincidence...). LLM-notebook learned from some of NotebookLM's ideas as it evolved. Here is how the main capabilities compare:

Dimension	Gemini Notebook (formerly NotebookLM)	LLM-notebook
What it can take in	Files, documents, web URLs, YouTube, pasted text	All of the above, plus LLM chats and Agent sessions
Answers and sources	Yes: answers mark their sources	Yes: claims point back to the paragraph, and [N] jumps there highlighted
What the Agent can do	Answer and generate artifacts inside the notebook	A native Agent that can read and write your material: edit a document, create a new one, organize and archive
What the output is	Audio and video overviews, mind maps, reports, flashcards, quizzes, infographics, presentations, data tables — made for consumption	Notes, mind maps, PPT — made for more work: editable, exportable
Access from China	Unreliable; often needs a VPN	Reachable directly, with no extra network setup

Try it, and tell us what you think

Project home: llm-notebook.com

Sign up with Google, WeChat or email (registration page). New users get a free 14-day trial plus $0.69 of token credit.

If you run into any problem while using it, you can submit feedback from Account → Feedback, or contact us directly by email at llm-notebook@intlinkbridge.com. Thanks for your support and feedback!