Open Notebook

Need the Open Notebook source and Docker files

Clone the repository locally

You obtain the Docker Compose files and app code needed to start the service

Need to use a live web article as a source

The page is uploaded and searchable

You can pull in web pages so the AI can answer from live articles or reports

Need a quick literature review across several PDFs

Get a comparative summary with citations

You can quickly generate a comparative summary of several PDFs, with citations to verify each point

Need separate notebooks for each research topic

Three named notebooks appear in the sidebar

Separate subjects into their own notebooks so you can switch context instantly

Organise existing documents

Move PDFs and web pages

Placing files in the right notebook keeps information grouped by relevance

Want to turn a webpage into a searchable note

Page becomes an indexed document

You can turn any online resource into a searchable document without manual copy-paste

Audio recordings aren’t searchable

They become searchable transcripts

Audio content becomes searchable text, expanding your knowledge base beyond PDFs

How does a concept appear across different notebooks?

You can discover how a concept appears across different topics and compare contexts

Start a podcast with all source docs ready

Preloaded notebook

You can start a podcast project by opening a notebook that contains the source documents

Want different voices for each line

Define up to four named speakers

You can define up to four distinct speakers so each line sounds like a different person

Need a consistent voice for your podcast

Choose academic, casual or debate tone

The overall style of the podcast can be set to match your audience

Can’t tell which notebooks to use

You control exactly which notes feed the script and audio

You control exactly which notes feed into the script and audio

Want to listen to a podcast

Playback or download MP3

You can immediately review the result in your browser or save it locally

Two papers clash

Instant debate‑style podcast

You can showcase opposing viewpoints by feeding two contradictory documents and selecting the debate tone

Want to see how tone changes your text

Side‑by‑side academic and debate versions

Switching tones lets you see how style affects the representation of content

Need a place to host a shared notebook app — get a VPS

You obtain a remote machine where the app can run and be reachable by your team

Need the codebase on your VPS

Cloned repository copy

You get the codebase needed to launch the application

Missing AI credentials

A .env file filled with your API key or Ollama settings

Open Notebook can call external AI APIs or a local Ollama server once it knows the key or endpoint

External users can't reach my notebook

Open port 3000 in firewall

Network traffic on the Open Notebook port can reach the server

Can’t open your notebook on other computers

View the web UI in any browser

Team members can interact with the app via a standard web browser

Custom domain with HTTPS

Secure address for your deployment

A friendly domain with TLS encryption makes access safer and easier

Your personal notebooks live only on your PC

They’re all moved to a shared server

Your team's collective notes become available in a central location

Can my team reach the shared notebook server?

A simple link confirms that access permissions and networking are correctly configured

How can I add an online article so the AI can use it as a source?

In any notebook, click ‘Add Source’ and choose ‘Web Page’. Paste the URL of the article and confirm; the page is fetched, its text is indexed like a PDF, and you can search it later.

What’s the best way to keep my PDFs and web pages organized by research topic?

Create separate notebooks for each topic (e.g., ‘RNA‑seq methods’, ‘Grant proposals’) and drag or upload the relevant files into them. If a document belongs to more than one topic, add it to each notebook so it appears in both contexts.

Can I make audio recordings searchable within Open Notebook?

Yes. Upload an MP3 or WAV file to a notebook and the system will generate a transcript, turning the spoken content into searchable text alongside your PDFs and web pages.

How do I quickly get a comparative summary of several papers?

Create a new notebook (for example, ‘Literature Review’), upload the three PDFs you want to compare, then ask a question like “What are the three main findings across these papers about [topic]?” The AI will synthesize the answers and provide citations you can click to verify each claim.

If I search for a term that appears in multiple notebooks, how are the results shown?

The search returns results grouped under each notebook where the term is found, letting you see how the concept is discussed in different contexts without mixing unrelated material.

No API keys set for your AI models

All providers connected and ready

The platform supports multiple LLM, embedding, TTS, and STT providers. Adding your API keys lets the notebook route requests to the chosen models, unlocking full functionality like chat, summarization, and podcast generation.

My files are all over the place

Get a chat‑searchable knowledge base that cites sources

You can upload files, paste URLs, or enter raw text as sources. The notebook indexes their content, allowing you to ask natural‑language questions that are answered using only the provided material.

Need a podcast from my notes

Get a downloadable audio file

Using built‑in podcast templates, Open Notebook can synthesize audio by feeding selected sources into TTS models, producing multi‑speaker episodes. This automates content creation for tutorials, interviews, or summaries.

Need a private AI notebook on your computer

Get Open Notebook running locally at http://localhost:8502

The video shows how to pull the Open Notebook Docker image, provide an OpenAI API key, and start the container. This gives you a self‑hosted instance that runs on your machine without needing to compile code.

Want to pull in PDFs, web pages, etc. as data sources

They become embedded and queryable

After the server is running, you can create data sources by uploading files or providing URLs. The tool embeds these sources into a vector store so they can be queried later.

Need to select embedding and chat models for a notebook

Queries answer correctly

The platform requires you to select both an embedding model (for vectorizing sources) and a chat model (for answering queries). Selecting the correct models in the Settings panel enables context‑aware conversations.

Need a private research‑chat notebook like Google’s

Run Open Notebook locally in the browser

Open Notebook is an open‑source, Docker‑first application that provides the same research‑chat experience as Google NotebookLM but runs on your own server. By running it locally you keep all uploaded documents private and can choose any LLM provider.

Need automatic research summaries from documents

Get concise JSON output

Open Notebook exposes a REST API that lets you programmatically add sources, trigger summarisation, and retrieve answers. This enables integration with CI pipelines, Slack bots, or custom dashboards without manual UI interaction.

Want a spoken version of any doc

Get a custom AI‑generated podcast with chosen voices

The built‑in podcast generator can turn any uploaded source into an audio narration using selectable speaker profiles. By configuring multiple speakers you can simulate a dialogue (e.g., product manager vs. backend engineer) to make dense material more engaging.

Want lower cost or keep data private

Notebook uses a different LLM backend

Open Notebook abstracts the model layer, allowing you to point it at hosted APIs (e.g., OpenAI, Anthropic) or local models served via Ollama. Changing the provider lets you balance response quality, latency, and data sovereignty.

Need a brand‑new AI‑assisted notebook for your project files

A ready‑to‑use Copilot Notebook

You open the Microsoft 365 Copilot experience, select the Notebooks tab and click “Create copilot notebook”. Naming the notebook gives you a dedicated workspace where all added content is linked, not copied.

Can't find my Word docs inside a notebook

Linked references to any OneDrive or Teams file

Within a notebook you can search or browse OneDrive, Teams channels, and recent files. Selecting a file creates a sharing link that stays in its original location while being accessible from the notebook.

Want each notebook to answer in its own style

Set custom Copilot instructions per notebook

Each notebook can store its own set of behavior guidelines. By adding “Copilot instructions”, you tell the AI how to respond (tone, format, focus) for all queries in that notebook.

How do I get answers only from my notebook's linked files?

When you type a query in the notebook’s chat, Copilot searches exclusively the linked references and notes inside that notebook, ignoring external data. This yields concise answers tailored to your project.

Need a fast way to review notebook notes

Get an AI‑generated audio summary

After gathering all relevant documents, you can click “Create audio overview”. Copilot synthesizes a spoken summary that reads like a podcast, allowing quick auditory review of the project.

Want a local AI notebook but hate command‑line installs

Get Open NotebookLM running in minutes

Claude Code provides an interactive terminal that automates the installation of Docker, pulls the Open NotebookLM repository and configures it for you. This removes manual command‑line steps and lets you get a running notebook in minutes.

Need Open NotebookLM but don’t want to install anything

Run it in an isolated container

Docker packages all dependencies into an isolated container, so you don’t need to install Python libraries or manage system paths. The video shows pulling the pre‑built image and running it with a single command.

Want a free local Llama model in Open NotebookLM

Appears as a provider and generates responses

Olama offers open‑source Llama models that can run on your CPU/GPU without API costs. The video demonstrates selecting Olama in Claude Code, letting the script download the Gemma 34B model and register it with Open NotebookLM.

Can't find relevant passages in my notebook sources

Semantic search using embeddings

Embedding models turn text into vectors that enable fast similarity search across added sources. The video adds an XY embedding tool via the settings UI, allowing notebooks to retrieve relevant passages from uploaded documents or web pages.

Want a private notebook server on your VPS

Get Open Notebook’s web UI running

Deploying Open Notebook via Docker on a virtual server gives you full control and privacy over your data. By pulling the GitHub repository into a Docker compose setup, the service runs in an isolated container that can be started with a single command.

Can’t get the notebook to talk to an AI model

It runs with your own API key

Open Notebook separates code from secrets; you must edit its configuration file to insert your own API key, otherwise the container cannot call external models. The YAML editor lets you replace placeholder values with real credentials safely.

Not sure which AI model to use in your notebook

Enabled models appear in the dropdown

After configuring the API key, you must activate specific model endpoints (language, embed, speech) in the Settings → Models page. This tells Open Notebook which provider and model to use for each task, allowing customized private assistants.

I want a personal notebook for my files and notes

AI answers using only my uploaded content

Notebooks let you upload files, links, or plain text as source material and add personal annotations. The system then uses only this curated data for answering queries, giving highly specific results while keeping everything on your VPS.

Want to keep AI provider keys safe in a notebook

YAML config stores them securely and adds the models

Open Notebook stores API credentials inside its Docker container's YAML configuration to avoid leaks. By editing the YAML file you can securely add keys for OpenAI or other providers, enabling chat and embedding models.

Private research space with sources and notes

AI answers from just that notebook

Each notebook isolates its data, so adding URLs, files, or plain text as sources keeps information scoped to that topic. Notes let you annotate findings, while the chat pane lets the selected model answer questions using only that notebook’s content.

Want to make your notes searchable and audible

Add embeddings, transcription and podcast creation

By configuring embedding models and audio models in the YAML file, Open Notebook unlocks vector search, transcription, and text‑to‑speech capabilities, allowing you to turn research notes into podcasts or perform semantic searches.

Want PDFs, docs, webpages and videos in one notebook

All sources ready for AI

You can import PDFs, Google Drive documents, websites, and YouTube videos into a notebook as sources. This lets the AI read and reference all of them when answering questions.

How do I get a summarized answer with citations from my docs?

The chat panel can query all added sources at once, and the response includes numbered citations that link back to the exact source content.

Want to keep an AI answer for later reference

It’s saved as a permanent note

After generating an answer, you can click Save as note to store it on the right‑hand side, making it easy to collect key insights for later reference.

I can’t search my own notes

They become AI‑searchable sources

User‑written notes can be converted into source entries, allowing the AI to treat them like any other imported document when answering queries.

My notebook is a mess of PDFs and web articles

They’re auto‑tagged into clear groups

Clicking the auto‑tag icon lets NotebookLM analyze your imported files and automatically assign them to logical groups, keeping large notebooks organized.

Want to see how ideas link across several docs

Get a visual mind map

The Studio can create visual mind maps that connect ideas across chosen sources, helping you see relationships without reading every document.

Need a spoken recap of your sources

Get an AI‑generated audio summary

Using the Studio’s audio option, you can produce a podcast‑style summary of selected sources, choosing language, length, and focus areas.

Need a private, offline AI notebook for documents

Get a self‑hosted Open NotebookLM

Open NotebookLM is an open‑source web interface that lets you upload documents, create notebooks and query them with LLMs running on your own machine. Because everything runs inside Docker it stays offline, giving full privacy and no usage limits.

No GPU but need a local LLM

Run Gemma 3 1B on CPU and get instant replies

Llama.cpp compiles LLM weights into an efficient CPU‑only binary, allowing you to run models like Gemma 3 1B without a GPU. The tool works with Open NotebookLM as the backend model for chat and tools.

Need keyword‑free searching of notes

Get fast semantic search across your knowledge base

An embedding model converts text chunks into vector representations that can be compared for similarity, powering fast keyword‑free searches across your knowledge base.

Turn my notebook notes into a ready‑to‑play podcast — MP3 file

Open NotebookLM can send text to a TTS service (e.g., ElevenLabs) and stitch the audio into a podcast, letting you repurpose notes or research as spoken media without leaving the platform.

Want to run Open Notebook without any installs

A ready‑to‑go isolated container

Docker packages all dependencies of Open Notebook into a self‑contained image, letting you run the app on any machine without manual installs. This isolation ensures consistent behavior and easy cleanup.

Want the full Open Notebook code on your machine

A complete cloned repository folder

Cloning the repository copies all required files to your local machine, giving you a complete, version‑controlled copy of the open‑source project ready for configuration.

Open Notebook can’t find my server or API keys

Custom .env with SERVER_IP and credentials

The .env file supplies environment variables that tell Docker where the app lives, which port to use, and which AI provider credentials to load. Editing it tailors the deployment to your network and chosen model.

Need a private AI notebook on your server

Get a live Open Notebook web UI

`docker compose up -d` reads docker‑compose.yml and the .env variables, builds any needed images, and starts the services in detached mode, making the web UI available on your network.