I cancelled Raindrop Pro after about a year. It’s a perfectly nice bookmark manager, and it never broke. But I was paying a separate subscription for something that should just live next to my notes, and my notes live in Notion.

I’m not really a developer. I decided where things should live and how the pieces connect. Claude Code wrote almost all the code. You don’t need to learn the Notion API or N8N to do this. You need to describe what you want and read the result.

There were two jobs: move everything I already had, then build a new way to save links.

Moving 6,400 bookmarks

When Pocket shut down, I’d dumped my whole archive into Raindrop: 5,256 bookmarks, most of which I’d never reopened. On top of that sat a few hundred useful links and around 200 tags, most of them subreddit names from the Pocket import.

I bought a Link Manager template for Notion, which gave me a ready-made structure instead of designing one. It splits links across four linked databases: Links, Folders, Tags and Categories. Raindrop is flat: each bookmark has a URL, a folder and some tags. So the script had to create the Folder and Tag rows first, then add each link pointing back to them.

I described that mapping to Claude Code in plain English, and it wrote a Python script to do it. The coding wasn’t the work. The work was saying “this becomes that.”

Then I hit Notion’s rate limit. The Notion API allows about 3 requests per second, and Notion doesn’t tell you that anywhere obvious. Each link took several calls (one to create it, more to connect its folder and tags). The script backed off and retried whenever Notion refused. The run took about four hours. I let it churn while I did other things.

Result: 6,140 links imported, 227 skipped as duplicates or empty, zero errors. I archived the Pocket pile in Notion so it stays searchable without cluttering the main view.

That afternoon, I cancelled Raindrop Pro.

I wanted saving to feel like it did before: tap Share in Safari, tap once more, done. No new app, no sorting afterward.

Here’s what I ended up with:

Safari Share Sheet
  → Apple Shortcut (sends URL to a webhook)
  → N8N workflow on my own server
      → fetch the page
      → clean the HTML
      → ask Claude Haiku for a clean title, a one-sentence description, and 1-3 tags
      → write it to the Notion Links database

Within about ten seconds the bookmark shows up in Notion. It has a clean title, a one-sentence summary and a couple of tags I would have picked myself.

The workflow has eight steps, and only one uses AI. I wrote about why I keep AI steps that small in an earlier post about giving Claude a server. If Haiku suggests a tag that doesn’t exist in Notion, the next step drops it. If the AI call fails, the bookmark still saves, just without the summary.

Claude Code built the whole workflow. It connects to N8N through an MCP server (a connector that lets Claude operate the tool directly). So it could deploy the workflow, run a test, read the error, fix it and try again. I watched.

Two things cost me time:

Apple Shortcuts hides things. “Receive from Share Sheet” isn’t an action you drag in. It’s a toggle in the Details panel that doesn’t look like a toggle. And “Get Dictionary Value” has a Type field that defaults wrong. If you compare the value as a string in an If action, set the type to Text, or the comparison silently fails.

My N8N server blocks $env in expressions for security reasons. The fix is to set up HTTP Header Auth credentials in N8N and use those instead. The error message doesn’t tell you that.

What it cost

  • A Notion template, one time (a few bucks, I don’t remember exactly)
  • Roughly a penny per saved bookmark for Haiku, rounded up
  • A few hours of my own time, mostly describing things and watching Claude Code work

And I save $29.96 a year on Raindrop going forward.

The AI-written summary turns out to be useful when I’m scrolling old saves and trying to remember why I kept something. I wanted my bookmarks to live next to my notes. Now they do, and I don’t think about them anymore.