Lapis0x0/obsidian-yolo

Agent-native AI assistant — chat, write, learning, all in one.

What it solves

YOLO is an agent-native AI assistant for Obsidian that transforms the note-taking app into a collaborative workspace. It solves the problem of fragmented AI workflows by integrating chat, writing assistance, and knowledge base management directly into the Obsidian vault, allowing users to interact with their personal notes using LLMs.

How it works

The plugin integrates with various LLM providers (OpenAI, Claude, Gemini, DeepSeek) and provides a set of agentic capabilities. It uses a hybrid RAG (Retrieval-Augmented Generation) system to ground answers in the user's vault notes. On desktop, it can interface with local CLI agents like Claude Code or Codex. It also supports the Model Context Protocol (MCP) and custom "Skills" to perform tasks and call tools.

Who it’s for

Obsidian users who want to integrate advanced AI agents, RAG-based knowledge retrieval, and structured learning systems into their personal knowledge management (PKM) system.

Highlights

  • Agentic Workflow: Supports tool calling, MCP servers, subagents, and desktop Bash access to move beyond simple Q&A.
  • Vault RAG: Uses local embedding models and multiple independent knowledge bases to retrieve information from notes without requiring an API key for indexing.
  • Learning Mode: Converts reference materials into structured learning projects with flashcards and FSRS-powered spaced repetition.
  • CLI Integration: Allows users to drive local CLI agents (Claude Code, Codex, Hermes, Pi) directly from the Obsidian chat interface.
  • Writing Tools: Includes real-time tab completion, "Quick Ask" editor integration, and cursor chat for seamless AI-assisted writing.

Related

  • Project
  • Project
  • Project
  • Project
  • Project