Obsidian Knowledge Tree

中文版

This script generates a knowledge tree for Obsidian by retrieving files with specific tags from your vault and using large language models to create a comprehensive summary.

This feature should ideally be built into an Obsidian AI plugin, but currently I haven't found a suitable one. Since I'm not familiar with plugin development or frontend work, I implemented this functionality as a Python script. PRs are welcome!

Usage Guide

1. First, install the Obsidian Local REST API plugin.

   • Enable the plugin
   • Allow HTTP service in the settings
   • 

2. Note down your Obsidian API key.

   • 

3. Clone this repository (or download the ZIP file) locally, then install dependencies:

   pip install -r requirements.txt

4. Configure the Obsidian API and your LLM API by creating a .env file with these variables:

   OB_API_BASE_URL="your_obsidian_api_url"
   OB_API_KEY="your_obsidian_api_key"
   LLM_API_BASE_URL="your_llm_api_url"
   LLM_API_KEY="your_llm_api_key"

5. Run the script:

Windows:

python obsidian_summarizer.py `
     --tag '#your_tag' `
     --model 'model_name' `
     --output 'output.md'

Linux/macOS:

python obsidian_summarizer.py \
     --tag '#your_tag' \
     --model 'model_name' \
     --output 'output.md'

You'll then receive an AI-generated summary document.

Follow me on WeixinMP/SSPAI (ID: 西郊次生林).