Open Notebook Integration: Connecting OpenClaw Agents with Local AI Research

What This Integration Does

This skill connects your OpenClaw agents to open-notebook, a powerful local AI
research assistant that serves as a free alternative to cloud-based notebook
solutions. The integration enables your agents to create thematic notebooks
for different purposes, save insights across sessions, and query accumulated
knowledge using local Ollama models.

Think of it as giving your agents a "second brain" that persists beyond
individual conversations. Whether you're conducting research, exploring agent
behaviors, or building personal knowledge systems, this integration provides
the infrastructure for long-term learning and discovery.

Core Benefits

The integration offers several compelling advantages. First, it eliminates API
costs by using local Ollama models, making it completely free to operate.
Second, it provides persistent knowledge storage, allowing agents to build
upon previous insights rather than starting fresh each time. Third, it enables
sophisticated querying capabilities, letting you ask complex questions about
your accumulated knowledge. Finally, it keeps everything local, ensuring
privacy and data sovereignty.

Prerequisites

Before you can use this skill, you'll need to set up the local infrastructure.
The system requires Docker Desktop for running open-notebook containers,
Ollama for local model serving, and at least one language model installed. The
setup process is straightforward but essential for the integration to
function.

Start by installing Docker Desktop from the official website. This provides
the container runtime that open-notebook needs. Next, install Ollama and
download a model - the qwen3-4b-thinking-32k model works well for most use
cases. Finally, run open-notebook using the provided Docker Compose
configuration.

Running open-notebook

You have two options for starting open-notebook. The simplest approach uses
the default configuration:

docker compose up -d

This launches both the SurrealDB database and the open-notebook application
with default settings.

If you want to use Ollama for model serving, use the host-specific
configuration:

docker compose -f docker-compose-host-ollama.yml up -d

This connects open-notebook directly to your local Ollama instance, ensuring
all processing happens on your machine.

Understanding the API Endpoints

Once running, open-notebook provides two main endpoints. The web interface is
available at http://localhost:8502, where you can create notebooks through a
graphical interface. The API server runs at http://localhost:5055, which is
what the PowerShell functions use to interact programmatically.

The API provides three main operations: creating new notebooks, adding content
to existing notebooks, and searching through notebook contents. These
operations are exposed through PowerShell functions that handle the HTTP
communication and JSON formatting automatically.

Available PowerShell Functions

The skill includes three PowerShell functions that you can use directly in
your scripts. The Add-ToNotebook function takes content and a notebook ID,
then sends the content to open-notebook for storage. The Search-Notebook
function allows you to query the contents of a specific notebook using natural
language questions. The New-Notebook function creates new thematic notebooks
with custom names and descriptions.

Each function is designed to be simple and intuitive. They handle the JSON
serialization, HTTP requests, and error handling internally, so you can focus
on the content rather than the mechanics of API communication.

Setting Up Notebook Identifiers

After creating notebooks through the API or web interface, you'll need to
capture their unique identifiers. These IDs are essential for the functions to
know which notebook to interact with. The skill provides several predefined
variables for common use cases: SIMULATION, CONSCIOUSNESS, ENJAMBRE, OSIRIS,
and RESEARCH.

You should update these variables with your actual notebook IDs after creating
them. For example, if you create a research notebook, you'd set $RESEARCH to
"notebook:your-research-id" and use that variable in your Add-ToNotebook and
Search-Notebook calls.

Practical Usage Examples

Using the integration is straightforward once your environment is set up. To
create a new notebook for research, you'd call:

New-Notebook -Name "My Research" -Description "Research notes"

Then, to save insights to that notebook:

Add-ToNotebook -Content "This is my insight" -NotebookId "notebook:xxx"

Finally, to query what you've learned:

$result = Search-Notebook -Query "What did I learn about X?" -NotebookId "notebook:xxx"

These simple commands provide powerful knowledge management capabilities for
your agents.

Configuration and Troubleshooting

Before using the functions, you must complete several configuration steps.
First, ensure open-notebook is running with Docker. Second, create at least
one notebook through the API or web interface. Third, capture the notebook ID
from the API response. Fourth, update the notebook ID parameters in your
scripts.

If you encounter issues, several troubleshooting steps can help. Check that
all containers are running with docker ps. Examine open-notebook logs with
docker compose logs. Verify Ollama is accessible with curl
http://localhost:11434/api/tags. These commands help isolate whether the
problem is with Docker, open-notebook, or Ollama.

System Requirements

The integration has specific system requirements to function properly. You
need Docker Desktop running to manage the containers. Ollama must be installed
with at least one language model available. The open-notebook containers must
be running, including both the SurrealDB database and the application itself.

Storage requirements depend on your usage patterns. Each notebook stores text
content, so the primary space consideration is the language models you choose
to install with Ollama. The system is designed to be lightweight and
efficient, making it suitable for personal computers and development
environments.

Version History and Updates

The current version is 1.0.1, which includes improved documentation and
comprehensive function examples. The development team continues to refine the
integration based on user feedback, with future versions likely to include
enhanced error handling, additional configuration options, and expanded
functionality.

The skill represents a mature integration that balances simplicity with
powerful capabilities. Whether you're a researcher, developer, or AI
enthusiast, this tool provides a solid foundation for building intelligent
agents with persistent knowledge capabilities.

Skill can be found at:
integration/SKILL.md>