正在加载视频...
视频加载失败
CLAUDE CODE + OBSIDIAN = KARPATHY'S SECOND BRAIN, BUILT FROM 7 FREE PLUGINS Andrej Karpathy: "Obsidian is the IDE; the LLM is the programmer; the wiki is the codebase." these 7 free plugins put Claude Code on Opus 5.5 inside that IDE. 38,457 GitHub stars, and your Obsidian vault... show more
23 条评论
building the whole list as an actual obsidian note with the graph view lighting up each plugin, Buzz that is next level
Do these plugins work with Claude Code out of the box?
claudian plus local git history is all you need here
7 plugins for a second brain is how you get a second junk drawer. Storage was never the bottleneck. Deciding what's worth remembering is.
which one stays on?
does smart-connections actually add much once claude code lives in the vault? your own memorytax point kinda says plain search already wins
acoustic version at open mic quality with us
Hi, please check your DM — I have a collaboration opportunity for you.
Giving an agent your whole vault is powerful. I'd still keep private notes out of its working folder.
Before you let an agent write into your vault: put the folder in git, and give the agent its own subfolder for its pages. Agents rename and rewrite in bulk; a diff after each session shows what changed, and one command undoes it. I make Rho MD, a free app for .md note folders.
#3 for me. a vault that knows everything about you is the last thing i'd want calling an api
The plugins are cool but the real trick is the wiki part. My bot writes its own notes after every task and reads them first next time. Way fewer repeat mistakes, no plugins needed
Curious which plugin handles large vaults best
yeah karpathy always knows the deal man
Obsidian as the IDE and Claude as the programmer is the cleanest framing I’ve seen. Bookmarking the plugin stack.
Letting the agent use the vault as its working folder is great until it quietly rewrites a note you needed. Putting the vault under git and reviewing the diff before committing makes it safe. Do any of the 7 handle versioning on their own?
the rest api one is where i'd be careful, any mcp client holding that key gets write access across the whole vault
Karpathy's line only works if the wiki gets treated like code too: put the vault in git, let Claude Code write into one inbox folder, and review the diff before it lands in your real notes. Otherwise 7 plugins later you have a second brain full of confident AI notes you never read.
7 free plugins, then somebody’s Obsidian vault becomes a dependency graph with feelings
Seven plugins is impressive, but I’m more interested in the maintenance side. Once a vault gets messy and interconnected, does Claude Code still handle it cleanly? @shuizhuyu
Karpathy's framing works because the wiki is just markdown, so the agent can read and edit it like code. If you could only keep one of the 7, which one?
Obsidian as the IDE is smart. Real second brain, or another plugin rabbit hole?
putting the agent inside the vault is fine... it still starts blank if you open a different tool its not anymore memory inside just agents but every other tool that you use as well
![Obsidian + Claude Code turned a 4,300-note vault into a second brain that answers back. Most people use Obsidian as a graveyard. You save 40 highlights a week, link nothing, and reread maybe 2% of it. The knowledge is there the retrieval is dead. Claude Code fixes the retrieval, because your vault is just a folder of markdown files and Claude Code lives in folders. The whole setup takes 20 minutes: Open a terminal inside your Obsidian folder and run Claude Code no plugins, no API glue, no export step. It reads all 4,300 files natively. Then drop a CLAUDE.md at the vault root explaining your structure: where daily notes live, how you tag, what your MOCs are. Now every session starts with context instead of chaos. From there you just talk to your vault: > "find every note where I mentioned churn and write a summary with backlinks" > "read my last 30 daily notes and list the 5 ideas I keep circling but never ship" > "build a MOC for everything tagged startup-ideas and link the orphan notes into it" It writes new notes straight into the vault with proper [[wikilinks]], and they show up in your graph view 10 seconds later. The boring layer automates too one command cleans broken links, one merges duplicates, one turns 6 months of meeting notes into a single decisions log. Work that used to eat a Sunday now runs while you make coffee. Before: 4,300 notes, 900 orphans, search by keyword and prayer. After: an agent that read everything you ever wrote and drafts in your own voice, from your own sources, with receipts. People pay $30/month for AI note apps with 10% of this. You already own the other 90%. Your notes stopped being storage. They started being staff.](https://image.24vids.com/tw-2078404662626074695/amplify_video_thumb/2078389656312217600/img/htCh1Y9vO-Wt4ERa.jpg)
