How to connect your bookmarks and notes to ChatGPT and Claude

The real power of an AI second brain is not just storing information — it is being able to query it conversationally. By connecting your CapturedIt library to ChatGPT and Claude, you transform static bookmarks into an interactive knowledge base.

Why connect your library to AI assistants?

Standard AI assistants have a knowledge cutoff and no access to your personal data. By connecting your library:

• Get answers grounded in sources you trust
• Reference specific research without searching
• Synthesize insights across months of accumulated knowledge
• Ask follow-up questions that drill into your saved content

How it works: MCP (Model Context Protocol)

MCP is an emerging open standard that allows AI assistants to connect to external data sources. Think of it like a USB port for AI — standardized, extensible, and increasingly supported by major assistants.

With MCP, your AI assistant can:

• Search your library for relevant items
• Retrieve full content for context
• Save new items back to your library
• Answer questions based on your accumulated knowledge

Setting up ChatGPT integration

Prerequisites

• CapturedIt Vault plan (required for MCP access)
• ChatGPT Plus or Pro subscription
• Your CapturedIt API key

Steps

1. Get your API key: In CapturedIt, go to Dashboard → Connections → API and generate a key.
2. Create a Custom GPT: In ChatGPT, go to Explore → Create a GPT.
3. Configure the action: Add an action using our OpenAPI spec. Paste the schema URL: https://api.capturedit.app/api/openapi.json
4. Authenticate: Set the API key in the authentication settings.
5. Start querying: Your Custom GPT can now search and reference your CapturedIt library.

Setting up Claude integration

Option A: Claude.ai (web)

1. Get your API key: Generate in Dashboard → Connections → API.
2. Add MCP server: In Claude.ai settings, go to Connectors → Add custom connector.
3. Enter URL: https://api.capturedit.app/api/mcp
4. Sign in: Authenticate with your CapturedIt credentials.
5. Use tools: Ask Claude to search your library or get context.

Option B: Claude Desktop

1. Open Claude Desktop settings
2. Navigate to the MCP server configuration
3. Add the CapturedIt MCP server URL: https://api.capturedit.app/api/mcp
4. Restart Claude Desktop
5. Your library is now accessible via natural language

Setting up Cursor integration

Cursor supports MCP for both inline chat and composer. This is particularly powerful for developers who want to reference documentation and saved solutions.

1. Open Cursor settings
2. Go to Features → MCP
3. Add server URL: https://api.capturedit.app/api/mcp
4. Add your API key in the authentication field
5. Save and restart Cursor

Now in Cursor, you can ask: “Search my library for how I handled CORS in Next.js” and get relevant results from your saved bookmarks and notes.

Example queries to try

Security and privacy

Your data remains yours. MCP connections use your API key and authenticate securely. CapturedIt never trains AI models on your content, and you can revoke API access at any time from your dashboard.

Conclusion

Connecting your knowledge base to AI assistants transforms how you work. Instead of static bookmarks you occasionally revisit, you have an active research partner that knows everything you have saved and can help you synthesize, compare, and create.

MCP is the future of AI tool integration. By adopting it now, you are building workflows that will only become more powerful as the ecosystem grows.