Adore NotebookLM but dislike Google? Explore this open-source substitute!

Open Notebook closely resembles the primary features of NotebookLM, including highlights like audio summaries and the ability to inquire about your research sources. While the tool has impressive qualities, it also presents some limitations. Below, I’ll outline what I appreciate about Open Notebook, its drawbacks, and how you can experiment with it yourself.

The fundamental features mirror NotebookLM, letting users upload diverse sources from PDFs to videos, subsequently synthesizing this content into audio summaries, quizzes, or through interactive chats.

What distinguishes Open Notebook from NotebookLM is its open-source nature, which necessitates user configuration. Although this presents certain challenges, it guarantees more robust data security compared to NotebookLM. Additionally, users benefit from significant flexibility and customization options beyond what providers like Google offer.

Open Notebook can be operated locally via models like Ollama or connected to popular cloud AI models, including GPT and Gemini Pro. Switching between various models is straightforward, allowing users to opt for multiple choices.

Moreover, Open Notebook does not impose limits on the number of notebooks you can create or sources you can use, which is a stark contrast to the free version of NotebookLM, which restricts users to 100 notebooks with a maximum of 50 sources each. However, be mindful that using a cloud-based model could still drain your token limits.

Although I find many aspects of Open Notebook appealing, it’s crucial to acknowledge its limitations.

Primarily, it is not a straightforward plug-and-play solution. As someone experienced with tech, I faced numerous challenges installing Open Notebook manually on a Chromebook. After multiple attempts, I successfully configured it on Windows more easily using Docker Desktop, taking about half an hour, but it still required some technical know-how.

Additionally, the audio summary features fall short in comparison to NotebookLM. Although users can select from up to four voice options instead of just two, the default audio summaries are significantly shorter. While NotebookLM delivers audio summaries lasting 8-15 minutes, Open Notebook’s are generally just a few minutes long.

This can be customized to extend up to 30 minutes or more, but even at that length, I found NotebookLM’s audio quality to be superior.

Moreover, Open Notebook requires a core PC or cloud setup to operate, and it lacks a native application; users must rely on a web browser for access.

Installing Open Notebook can indeed be challenging, but the simplest method is using Docker Desktop, which allows for running containerized applications with minimal command-line interaction.

After installation, obtain an API key from your selected provider; the steps will vary depending on your choice.

Even though an official setup guide exists, it may contain outdated information. For those setting up a cloud model, the following steps provide a basic initial configuration.

Begin by creating a configuration file. Use a text editor to input the following code:

services:
surrealdb:
image: surrealdb/surrealdb:v2
command: start –user root –pass password –bind 0.0.0.0:8000 rocksdb:/mydata/mydatabase.db
ports:
– “8000:8000”
volumes:
– ./surreal_data:/mydata
restart: always

open_notebook:
image: lfnovo/open_notebook:v1-latest
pull_policy: always
ports:
– “8502:8502” # Web UI
– “5055:5055” # API
environment:
– OPEN_NOTEBOOK_ENCRYPTION_KEY=change-this-to-your-secret
– SURREAL_URL=ws://surrealdb:8000/rpc
– SURREAL_USER=root
– SURREAL_PASSWORD=password
– SURREAL_NAMESPACE=open_notebook
volumes:
– ./notebook_data:/app/data
depends_on:
– surrealdb
restart: always

After saving your file, replace change-this-to-your-secret with a preferred phrase and save it as docker-compose.yml in a new folder named Open Notebook.

Next, run Docker Desktop. Once operational, navigate to the Open Notebook folder using File Explorer, right-click the window, and select Open in terminal.

In the terminal, input: docker compose up -d.

After approximately 15 seconds, the services will commence, and you can access Open Notebook by entering http://localhost:8502 into your web browser. However, you are not finished yet.

Go to Manage > Models, select your chosen AI provider, and click the Add Configuration button. Enter a name for your configuration and the API key in the prompt that appears, and click Add Configuration.

Afterwards, use the Test button next to your configuration to check the connection. Specify which model you want for chat, embedding, and transformation tasks such as creating audio summaries, and more.

Once everything is set up, you can create a new notebook in the New tab. From there, add sources and explore the various transformation capabilities. You will need to configure your podcast settings and other minor details; however, I will refrain from elaborating on those for brevity, as more information is available in the Open Notebook guide.

The question remains: is Open Notebook a better option than NotebookLM? The answer varies depending on your individual needs and your willingness to fine-tune settings to suit your preferences. Open Notebook might be an appealing choice for those who appreciate NotebookLM but have reservations about data residing on Google servers. It also attracts those preferring an alternative model rather than one based on Gemini technology.

For others, NotebookLM remains a slightly more efficient and user-friendly choice, especially if enhancing features through NotebookLM Pro or Plus is of interest. However, if you’re comfortable with some adjustments, Open Notebook is definitely worth a try.