Saltar al contenido
TakoAPI
Volver a skills

open-notebook-integration

por nantes

Coding Agents & IDEs
425Descargas según clawskills.sh
2Estrellas
1Vistas
0 me gustaInicia sesión para dar me gusta a esta skill
—Aún sin valoraciones
Tu valoración

Resumen

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

Descripción

Qué hace esta skill

Connects OpenClaw agents to open-notebook, a locally hosted research assistant. Agents can create notebooks, save content, and query stored knowledge using PowerShell functions that call the open-notebook REST API. Runs fully local via Docker and Ollama with no external API costs.

Runs entirely on local infrastructure with Ollama, so there are no API costs or data privacy concerns compared to cloud-based alternatives like NotebookLM.

Cuándo usarla

  • Saving agent-generated research findings across sessions
  • Querying past notes to continue interrupted research
  • Organizing knowledge into separate thematic notebooks
  • Building a persistent memory layer for long-running agent workflows
  • Replacing cloud-based notebook tools with a self-hosted alternative

Flujo de trabajo de ejemplo

ENTRADA

User asks: Save my notes on transformer architectures and let me query them later

AGENTE
  1. 1

    Calls New-Notebook to create a notebook named 'Transformer Research'

  2. 2

    Records the returned notebook ID

  3. 3

    Calls Add-ToNotebook with the content and notebook ID to save the notes

  4. 4

    Confirms the source was added via the API response

  5. 5

    Calls Search-Notebook with a query to retrieve relevant stored knowledge

SALIDA

Notebook created and notes saved locally; agent returns the answer retrieved from stored content

Requisitos

  • Docker Desktop installed and running
  • Ollama installed with at least one model (e.g. qwen3-4b-thinking-32k)
  • open-notebook containers running locally via Docker Compose

Instalación

clawhub install nantes/open-notebook-integration

Ejecuta el comando de la CLI o pega el prompt en Claude Code, Cursor, Codex o el agente de programación que uses.

Enlaces