Everythings
A good conversation produces thinking worth keeping, and then it stays in a chat window that almost nobody scrolls back through. Everythings gives that output somewhere to live. A workspace holds a tree of "things". A thing is one unit of thought: a name, a Markdown body, and a position among other things. Things nest inside each other, so part of what a thing means comes from where it sits. This connector lets Claude read and write that structure directly, so a conversation becomes an input method and the thing is where the result stays. What Claude can do: - Search across every workspace you have access to, by name, tags, or body text - Read a thing's full content, along with its sub-items, comments, and reactions - Create and edit things, nest them, reorder them, and move or copy them between workspaces - Create and rename workspaces - Add comments and react with emoji marks - Ask you a question on a thing, and read your answer later Ask Claude to keep what you just worked out ("save this under Travel") or pull back something from months ago ("what did we decide about the roof?"). When Claude works on something while you are away, it can stop and ask. The question waits on the thing in the Everythings app, which notifies you; you answer from your phone or the web, and Claude reads the answer on its next run. Everything Claude writes is labeled with its name, and an inbox in each workspace lists what changed since you last looked. Read tools are annotated read-only, so they run without a confirmation each time. Edits and deletions are annotated destructive, so Claude asks before making them. Editing replaces content in full, so Claude reads the current version first. Authentication is OAuth. Sign in with your Everythings account, grant access once, and revoke it at any time from your Everythings settings. The connector reaches only the workspaces your account can open. An Everythings account is all you need; the free plan is enough.
Details
https://www.everythings.app/api/mcp · streamable-http
