Blog
The team second brain
On real-time collaboration, local-first markdown, open source, and giving your AI a second brain.
Page 4 of 7
Pick a second brain for Claude on the write path, not the read
Every second brain for Claude reads your notes well. They differ on what happens when Claude writes: whole-file overwrite, anchored edit, or a refusal.
Oct 10, 2026 · 9 min readAn MCP server for team notes has to know who is asking
Most note MCP servers authenticate with one key over one folder. What changes when the notes belong to a team, and what a server needs to handle it.
Oct 10, 2026 · 5 min readSelf-hosting Obsidian notes is three different jobs people keep confusing
Self-hosting Obsidian means three unrelated things: a sync backend, a desktop in a container, or a published site. What each gives you, and what none does.
Oct 10, 2026 · 9 min readWhy Obsidian sync conflicts happen, and why merging afterwards is too late
Obsidian Sync merges markdown with diff-match-patch once both copies are finished, which is why text comes back doubled. What removes the conflict instead.
Oct 10, 2026 · 8 min readGive an AI agent memory that a person can open and correct
How to give an AI agent memory: Mem0, Zep and AgentCore keep extracted records in their own store. For a team, memory has to be a file somebody can open.
Oct 10, 2026 · 10 min readCan Obsidian be a team knowledge base? Only if everyone reads everything
Obsidian Sync shares the whole vault with everyone. Publish locks a whole site behind one password. What a team knowledge base needs that neither can give.
Oct 10, 2026 · 9 min readMoving a team from Notion to Obsidian without losing the multiplayer half
How to move from Notion to Obsidian: the docs say skip Notion's Markdown export. The right one, what the Importer keeps, and where your collaborators go.
Oct 9, 2026 · 8 min readInternal documentation: write the decisions, index the files you already have
Most guides treat internal docs as a writing problem. Half of what your team knows is already a file nobody will retype. What to do with each half.
Oct 9, 2026 · 8 min readClaude Code does have memory now. It just never leaves your machine
Claude Code keeps two memory systems across sessions, both markdown on one laptop. What persists, where the files sit, and what a team has to do instead.
Oct 9, 2026 · 8 min readChatGPT makes a good second brain and a bad place to keep one
Custom instructions and Projects make ChatGPT behave like a second brain, inside your account. Connecting it to an MCP server makes it the interface to one.
Oct 9, 2026 · 8 min readConfluence alternatives that use markdown, and where each one actually puts the file
Confluence stores XHTML and never exports markdown. Which alternatives keep your markdown in a database, and which write real files to your disk.
Oct 9, 2026 · 7 min readA Claude Code second brain for a team: what breaks when the second person arrives
The solo Claude Code second brain needs no setup at all. Here is what breaks the moment a second person needs the same notes, and what fixes each part.
Oct 9, 2026 · 7 min readStart your team’s brain
Free and open source. No account needed.