#1919th of 27 in Memory
obsidian-wiki
Markdown skills that let coding agents compile notes into an Obsidian wiki
- Stars
- 3.7k
- License
- MIT
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- Python
546 on 22 Apr 20263,718 on 11 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
obsidian-wiki is a set of 39 markdown skills, plus an optional Python CLI and library, that has agents such as Claude Code, Cursor and Codex ingest docs, chats and URLs into linked markdown pages in an Obsidian vault. It follows Karpathy's LLM Wiki pattern: merge knowledge into existing pages, tag claims as extracted or inferred, and answer queries with wikilink citations. Session hooks inject a recap at start and suggest a capture at end.
Who it is for: Developers who use coding agents and keep notes in Obsidian
Strengths
- Knowledge is plain markdown on disk, readable, greppable and diffable; no vector store or API key
- Manifest tracks ingested sources, so repeat runs process only changed files
- Works across many agents from one skills directory, including Claude Code, Cursor, Codex and Gemini CLI
- Lint, dedup and cross-linking skills flag contradictions, orphans and speculative pages
Weaknesses
- Quality depends on the host agent and model; the project ships no model of its own
- Audio and video are skipped, and Office file handling depends on the agent
- Benchmark is small: n=2 per cell on one 38-page vault
- Claude Code plugin install omits the session hooks and the CLI
What it needs
- no GPU
- Docker + Compose
- Needs Obsidian (optional viewer), An AI coding agent such as Claude Code, Cursor or Codex
- Models: Any model via the host agent
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in Memory
See all 27| Rank | Project | Score |
|---|---|---|
| 1 | hindsightAgent memory server with retain, recall and reflect operations | 95 out of 100 |
| 2 | gbrainMemory store with sources and a knowledge graph for coding agents | 91 out of 100 |
| 3 | MemPalaceLocal verbatim memory store for AI agents with semantic search and MCP | 90 out of 100 |
| #4 | SupermemoryMemory and context engine for AI agents, with user profiles and hybrid search | 89 out of 100 |
| #5 | Mem0Long-term memory layer that stores and retrieves user facts for LLM agents | 88 out of 100 |