Open Notebook Skill

Integrates OpenClaw agents with local open-notebook for creating, saving, and querying thematic notebooks using local Ollama AI models.

Open Notebook Integration

A skill for integrating OpenClaw agents with open-notebook, a local AI research assistant (NotebookLM alternative).

What It Does

• Connects your agent to open-notebook running locally
• Creates thematic notebooks for research, agent discovery, and personal knowledge
• Enables saving and querying knowledge across sessions (second brain for agents)
• Supports local Ollama models (free, no API costs)

Prerequisites

1. Install Docker Desktop (required for open-notebook)

2. Install Ollama with a model (e.g., qwen3-4b-thinking-32k)

3. Run open-notebook:

   powershell

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

   Or use the default compose:

   powershell

   docker compose up -d
   

Setup

The skill expects open-notebook at:

• UI: http://localhost:8502
• API: http://localhost:5055

Functions (INCLUDED)

This skill provides these PowerShell functions directly:

Add-ToNotebook

powershell

function Add-ToNotebook {
    param(
        [string]$Content,
        [string]$NotebookId = "YOUR_NOTEBOOK_ID"
    )
    $body = @{
        content = $Content
        notebook_id = $NotebookId
        type = "text"
    } | ConvertTo-Json
    Invoke-RestMethod -Uri "http://localhost:5055/api/sources/json" -Method Post -ContentType "application/json" -Body $body
}

Search-Notebook

powershell

function Search-Notebook {
    param(
        [string]$Query,
        [string]$NotebookId = "YOUR_NOTEBOOK_ID"
    )
    $body = @{
        question = $Query
        notebook_ids = @($NotebookId)
        strategy_model = "model:YOUR_MODEL_ID"
        answer_model = "model:YOUR_MODEL_ID"
        final_answer_model = "model:YOUR_MODEL_ID"
    } | ConvertTo-Json
    Invoke-RestMethod -Uri "http://localhost:5055/api/search/ask" -Method Post -ContentType "application/json" -Body $body
}

New-Notebook

powershell

function New-Notebook {
    param(
        [string]$Name,
        [string]$Description = ""
    )
    $body = @{
        name = $Name
        description = $Description
    } | ConvertTo-Json
    Invoke-RestMethod -Uri "http://localhost:5055/api/notebooks" -Method Post -ContentType "application/json" -Body $body
}

Notebook IDs

After creating notebooks, update these variables in your scripts:

powershell

$SIMULATION = "notebook:YOUR_SIMULATION_ID"
$CONSCIOUSNESS = "notebook:YOUR_CONSCIOUSNESS_ID"
$ENJAMBRE = "notebook:YOUR_ENJAMBRE_ID"
$OSIRIS = "notebook:YOUR_OSIRIS_ID"
$RESEARCH = "notebook:YOUR_RESEARCH_ID"

Example Usage

powershell

# Create a new notebook
New-Notebook -Name "My Research" -Description "Research notes"

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

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

Configuration Required

Before using, you MUST:

1. Run open-notebook with Docker
2. Create notebooks via the UI (http://localhost:8502) or API
3. Get your notebook IDs from the API response
4. Update the $NotebookId parameters in the functions

Requirements

• Docker Desktop running
• Ollama with at least one model installed
• open-notebook containers running (SurrealDB + app)

Troubleshooting

• If API fails, check containers: docker ps
• Check open-notebook logs: docker compose logs
• Verify Ollama is running: curl http://localhost:11434/api/tags

Version

1.0.1 - Improved documentation, included function examples