Aether as an MCP connector

Use your Aether knowledge base directly inside Claude and ChatGPT — no code required.

Aether is a managed knowledge base for AI. You save notes, documents, and references into it, and Aether finds the right passages by meaning whenever you (or an AI) ask a question. The Model Context Protocol (MCP) is an open standard that lets AI assistants connect to outside tools and data. Putting the two together means Claude and ChatGPT can search and update your Aether knowledge base from any conversation.

You connect once. After that, the assistant can pull facts out of your knowledge base — and save new ones back into it — without you copying, pasting, or re-uploading anything.

Who this is for

This is for anyone building with AI who wants knowledge that sticks around and travels with them:

• Vibe coders and makers who live inside Claude or ChatGPT and want the assistant to remember their project's decisions, specs, and notes.
• Business analysts and operators who keep a growing pile of research, meeting notes, and reference docs and want to ask questions across all of it in plain language.
• Builders using the SDKs who want the same knowledge base their app reads from to also be available inside their assistants.

You don't need to write any code to use the connector. If you later want to build on top of the same knowledge base, the Python, TypeScript, Go, and .NET SDKs read and write the exact same data.

Why a connector instead of uploading files

Most assistants let you attach a file to a chat. That works for a one-off question, but the file lives and dies with that conversation. Upload the same report into ten chats and you've uploaded it ten times.

Aether is different — it's a persistent, managed corpus that you build up once and reuse everywhere:

Because the corpus is managed centrally, the knowledge you save in Claude is instantly available in ChatGPT, in Claude Cowork, and to any app you build on the Aether API. One knowledge base, every surface.

How it works

When you add Aether as a connector, your assistant gets a small set of tools it can use on your behalf — searching your knowledge base, saving new documents, listing and managing what's there. You stay in control: you sign in to Aether through a secure browser flow and approve exactly which workspace the assistant may access.

You: "Save this summary to my Aether knowledge base"
        ↓
Assistant calls the Aether connector → document stored
        ↓
Later, in any chat: "What did we decide about pricing?"
        ↓
Assistant searches Aether → cites the saved passage → grounded answer

Authentication uses OAuth 2.1 — you sign in, you never paste an API key into the chat. Each connection is scoped to a single Aether workspace, and you can review or revoke it at any time. See Manage your connection & privacy for the details.

Get connected

Pick your assistant and follow the steps. The flow is the same on both: add the connector URL, sign in to Aether, approve access.