Day 20 - Migrating Twitter Bookmark Search to IndexedDB, and trying out Claude Code

About 2 years ago, I was frustrated that Twitter didn't have a native way to search your bookmarks there. So I built a tool that scrapes your Twitter bookmarks and uploads them to an app I built where you can search and filter everything. You can visit it at: https://www.twitter-bookmark-search.com/.

When you upload your bookmarks, I have the app save them to localstorage so that you don't have to upload them every time you visit the site. It had been a while since I'd last scraped them, and when I scraped them recently, I realized that it wasn't saving because it exceeded the localstorage size limit of 5mb. So I spent today migrating the storage layer to IndexedDB, which has a much larger capacity.

Separately, I also spent some time today playing around with claude code. I'd tried using it a bit in the past as a coding agent, but I didn't enjoy being in a cli environment when building software. It just didn't feel right. So I still use Cursor for coding use cases, but today I tried using Claude Code for the kinds of one-off tasks that involve the filesystem and benefit from running cli commands, like having it use ffmpeg to compress the images in a folder for example. It felt great, and maybe points to Claude Code's broader potential less as an agent just for writing code, and more as a general purpose computer-use agent.