sircharan/second-brain — explained in plain English
Analysis updated 2026-05-18
Keep Claude Code remembering project context across separate coding sessions.
Export a portable memory vault to use the same notes with ChatGPT or Gemini.
Avoid slow context compaction by clearing chats and recalling only relevant notes.
Search and audit an Obsidian-compatible vault of AI-captured session notes.
| sircharan/second-brain | 0xkinno/neuralvault | 0xmayurrr/ai-contractauditor | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | easy | hard | easy |
| Complexity | 2/5 | 4/5 | 2/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Installs via the Claude Code plugin marketplace or a bundled install.sh script, no external database needed.
second-brain is a file based memory system for Claude Code, the AI coding assistant. It quietly captures what happens in every coding session into a folder of plain Markdown notes that is compatible with the note taking app Obsidian, brings back the relevant notes into new prompts, and saves a snapshot of what you were doing right before the conversation's context window has to be trimmed. The goal is that Claude can pick up where you left off across different sessions, different projects, and even different machines. The project's core argument is about ownership. Most AI memory features keep your accumulated context locked inside one vendor's servers, so switching between ChatGPT, Claude, Gemini, or Grok means starting from zero each time. second-brain stores everything as plain Markdown files with wikilinks and YAML frontmatter on your own disk, with no database and no server, so you can read it with any tool, move it between assistants, or export the whole vault into one file to hand to a different model. Because memory lives in the vault rather than the chat window, you can clear a conversation and start fresh cheaply instead of waiting through a slow context compaction. Under the hood, each session start injects a short summary of where you left off along with relevant notes, each prompt pulls in matching notes, each reply logs a one line entry into a daily journal folder, and a consolidate step later distills that journal into curated, permanent notes. Old facts are never deleted outright, they are marked as retired and linked to whatever replaced them, so the whole history stays visible and auditable through git. It installs either as a Claude Code plugin through the plugin marketplace, or through a bundled install.sh script that copies hooks and configuration into your Claude Code settings folder. Once installed, a /second-brain skill provides commands for capturing facts, searching the vault, checking its health, exporting it, and more. There is also an optional MCP server that exposes the same vault to Claude Desktop, Cursor, and ChatGPT. The project is written in TypeScript with Python and bash hooks.
A file-based, vendor-independent memory system that gives Claude Code persistent notes across sessions and projects.
Mainly TypeScript. The stack also includes TypeScript, Python, Bash.
The README does not state a license.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
double-check against the repo, no cap.