How to Connect Glasp to Claude, ChatGPT, and Claude Code (MCP Connector)

🌐 Read this in your language: English · 日本語 · Español · Français · Português (BR) · Tiếng Việt · 한국어

Have you ever wished your AI assistant knew what you have been reading? With Glasp's MCP Connector, you can connect your Glasp account to any MCP-compatible AI tool, including Claude, ChatGPT, Claude Code, and clients built on the OpenAI API, and ask them about everything you have read, highlighted, and saved. Your AI assistant can search your highlights and memories in natural language and pull the exact passages you saved from articles, PDFs, YouTube videos, and Kindle books into any answer.

Picture this: instead of digging through old bookmarks, you ask Claude "what did I save about deep work?" and get your own highlighted passages back, with their sources, right inside the chat you already use.

The connector is read-only and private to you. It can read your own highlights and memories, and nothing else. It can never post, edit, or delete anything.

In this tutorial, we will walk you through how to connect Glasp to each of them: Claude, ChatGPT, Claude Code, and any other MCP-compatible client. The steps are almost the same everywhere, so once you have done one, the rest take a minute each.

Before you start

Go to the Settings page on Glasp and click MCP Connector in the left sidebar. You can also learn more on the MCP Connector page.

The MCP server URL you will use in every client is:

https://glasp.co/api/mcp

Connect your AI tool

Pick the tool you use below. You do not need to do all of them. The idea is the same everywhere: point the client at the Glasp server URL and authorize.

Claude

All you need is your Glasp account. The connection uses OAuth, so there is no token to copy or paste.

First, in Claude, open Settings, then Connectors.

Click Add, then Add custom connector. Enter a name (for example, Glasp) and paste the Glasp server URL: https://glasp.co/api/mcp. Then click Add.

Open the Glasp connector you just added and click Connect.

It opens Glasp's authorization page. Check that the access is read-only and click Authorize. That's it. Claude can now search and read your highlights and memories.

ChatGPT

You need a ChatGPT plan that supports custom connectors. Availability depends on your plan.

In ChatGPT, open Settings, then Connectors (or Apps).
Add a custom connector with the Glasp server URL: https://glasp.co/api/mcp
Sign in with Glasp and approve.

Claude Code

For Claude Code, you need a Glasp access token. You can create one at Settings then MCP Connector.

Create a token on the MCP Connector settings page.
Run the command below in your terminal, replacing YOUR_TOKEN with the token you created:

claude mcp add --transport http glasp https://glasp.co/api/mcp --header "Authorization: Bearer YOUR_TOKEN"

Ask Claude Code about your highlights.

Note: The token is shown only once, so please copy it right away. Tokens are stored only as hashes on Glasp's side.

The OpenAI API or any MCP client

This one is for developers and power users: if your tool has a place to add an MCP server, like Cursor, VS Code, or your own app built on the OpenAI API, these three steps are all you need.

The Glasp MCP Connector works with any client that supports the Model Context Protocol or the OpenAI API.

Create a token on the MCP Connector settings page.
Point your MCP client at the server URL: https://glasp.co/api/mcp
Send the token in the Authorization header: Authorization: Bearer YOUR_TOKEN

Ways to use it

Once connected, your highlights become something your AI can actually work with. A few ways people use it:

Write with your own sources, not generic AI text

Drafting an article or newsletter? Ask your assistant to pull the passages you already highlighted and weave them in with citations. You get writing grounded in what you actually read, not filler.

"I'm writing about deep work. Pull the quotes I highlighted on focus, and draft an intro that uses them."

Resurface what you forgot you saved

Highlighted something months ago and cannot find it? Just ask. It searches across every article, PDF, YouTube video, and Kindle book you saved.

"What have I saved about pricing strategy, and where was each from?"

Get a briefing before a meeting or decision

Turn scattered highlights into a one-paragraph brief.

"Summarize everything I've highlighted about a competitor or topic into five bullet points."

Review what you are learning

Turn your reading into study material.

"Quiz me on my recent highlights about AI agents," or "Summarize my Kindle highlights from a book."

Stay in your coding flow (Claude Code)

Ask about something you saved without leaving the terminal.

"What did I highlight about rate limiting in that API article?"

It's easy and useful, right? With the MCP Connector, your highlights and notes stop being a static archive and become a knowledge base your AI assistant can actually use. The more you highlight, the more your AI knows about what you have learned.

FAQs

Q. What is the Glasp MCP Connector?

It is a feature that connects your Glasp account to AI assistants such as Claude and ChatGPT as an MCP (Model Context Protocol) server. Your AI assistant can then search and read your highlights and memories to answer your questions.

Q. Is the MCP Connector free?

Yes. It is included with your Glasp account.

Q. What can the connector access?

Your own web highlights, Kindle highlights, and AI-generated memories. It is read-only, and every request is scoped to your account. It can never see another person's data.

Q. Which apps work with it?

Claude on the web, Claude Code, ChatGPT, and any client that supports the OpenAI API or the Model Context Protocol.

Q. Can it change or delete my data?

No. The connector is read-only by design and cannot create, edit, or delete anything.

Q. How do I disconnect it?

Open Settings, then MCP Connector to revoke tokens or disconnect connected apps at any time.

Before you leave

If you have any questions, please refer to the FAQ or contact us.

👉 FAQs

👉 X

👉 Reddit

See you next time,

Glasp team