odin  by memgraph

Obsidian plugin for LLM-powered knowledge graph exploration

created 1 year ago
595 stars

Top 55.5% on sourcepulse

GitHubView on GitHub
Project Summary

ODIN is an Obsidian plugin that transforms your notes into a navigable knowledge graph, powered by Large Language Models (LLMs). It's designed for Obsidian users who want to deepen their understanding of their personal knowledge base by uncovering thematic connections and generating insights.

How It Works

ODIN leverages LangChain to integrate LLMs, enabling users to query their Obsidian vault directly. It visualizes the entire vault as a dynamic graph, with options to view the whole vault or focus on specific files. The plugin analyzes highlighted text in the editor to suggest relevant links to other notes, generate insightful questions, and identify related nodes, thereby enriching the knowledge discovery process.

Quick Start & Requirements

  • Docker: docker compose up (requires Docker and Docker Compose, .env file with OPENAI_API_KEY and LLM_MODEL_NAME). Setup time: up to 10 minutes.
  • Manual: npm install then npm run dev (requires Node.js >= 14, npm). Also requires manual setup of Memgraph database, BOR backend, and Rune.
  • Obsidian: Must be installed with community plugins enabled.
  • OpenAI API: GPT-4 is strongly recommended; GPT-3.5 may produce incorrect knowledge graphs.

Highlighted Details

  • LLM-powered querying directly from Obsidian's prompt bar.
  • Dynamic knowledge graph visualization of the entire Obsidian vault.
  • Context-aware features: generate questions, predict links, and suggest nodes from highlighted text.
  • File-specific graph views with sentence highlighting.

Maintenance & Community

The project is maintained by Memgraph. Links to community channels or roadmaps are not provided in the README.

Licensing & Compatibility

The repository is licensed under the MIT License. This license permits commercial use and linking with closed-source projects.

Limitations & Caveats

The README states that the init_repo, update_file, and add_file endpoints are untested due to lack of direct GPT-4 API access during development. Manual initialization via ChatGPT was used. Users are warned "Here be dragons."

Health Check
Last commit

1 year ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
0
Star History
8 stars in the last 90 days

Explore Similar Projects

Starred by Chip Huyen Chip Huyen(Author of AI Engineering, Designing Machine Learning Systems), Eugene Yan Eugene Yan(AI Scientist at AWS), and
1 more.

obsidian-copilot by eugeneyan

0.2%
542
Prototype assistant for writing and thinking
created 2 years ago
updated 1 year ago
Feedback? Help us improve.