Connect your AI to the archive
The whole archive is also an MCP server. Connect it to Claude, ChatGPT, Cursor or any other MCP client, and your AI can search all 210 sources, read every essay and wiki page, and walk the knowledge graph, with links back to the pages here.
/mcpClaude
- Open Customize → Connectors, then Add → Add custom connector.
- Name it Lumberjack Archive, paste the address above, choose No sign in and add it.
- In a chat, open + → Connectors and switch it on.
Works on the web, desktop and mobile apps. The free plan allows one custom connector. On Team and Enterprise plans an owner adds it under Organization settings → Connectors.
Claude Code
claude mcp add --transport http lumberjack-archive /mcpChatGPT
In Settings → Apps & Connectors → Advanced settings, turn on Developer mode. Then choose Create, paste the address and set authentication to None. This needs a paid plan, and the menu names move around while the feature is in beta.
Cursor
Add this to ~/.cursor/mcp.json, or to .cursor/mcp.json in a project:
{
"mcpServers": {
"lumberjack-archive": {
"url": "/mcp"
}
}
}VS Code
Add this to .vscode/mcp.json:
{
"servers": {
"lumberjack-archive": {
"type": "http",
"url": "/mcp"
}
}
}Anything else
Any client that speaks MCP over Streamable HTTP can use the address as it is. The server is stateless, so there is no session to keep, and it speaks both the current and the older versions of the protocol.
What your AI can do with it
| Tool | What it does |
|---|---|
overview | What the archive is: the eight essays, every idea by theme, the projects and the kinds of sources. A good first call. |
search | Full-text search over essays, wiki pages and all 210 sources, filtered by type, theme, kind of source or date. |
read | Any essay, idea, project or source in full, as markdown, with its dates, related ideas and sources. Takes an id or a link to a page here. |
list | Browse one type: the essays, the ideas in a theme, the projects, or the sources by kind and date. |
neighbors | An idea's strongest links in the knowledge graph, and the sources it shows up in. |
path | How two ideas connect: the shortest chain of strong links, with the sources behind each step. |
graph | The knowledge graph itself, as nodes and edges: whole, by theme, around one idea, or as it stood in a given month. |
timeline | How the ideas grew month by month, or one idea's mentions over time. |
Things to ask
- Using the Lumberjack Archive, how did David's thinking about agents change between 2024 and 2026?
- What does David mean by the harness, and where did he first write about it?
- How does disposable software connect to glue people in the knowledge graph?
- Find everything David wrote about presence and screens, newest first, and quote him.
What's in it
Exactly what this site shows, and nothing more: the eight essays, 95 ideas, 11 projects, 5 pages that read across them, and 210 sources. Course lessons, emails, the textbook, workshop sessions and research appear as summaries, the way they do here. Recorded conversations aren't published. My thinking changed a lot between 2024 and 2026, so every page carries its dates, and the server asks your AI to say when a view is from.