Content Capture Bookmarklets • Ryan

I encounter interesting information in lots of different places - sometimes that’s from chatting with an LLM in a browser (often switching between different models or tiers), sometimes that’s a LinkedIn conversation or long-form article that I want to keep track of or write about, etc.

I’ve tried using systems like Evernote or Pinterest to capture and keep track of useful content, but they never worked long-term, mostly because it got to be too difficult/stressful to figure out on the spot how to find anything again later. I just ended up with lots of orphan folders and tags with one item in them.

I wanted something that:

saves content from different sources in a single, portable format
lives right where I’m already working (i.e., in my browser)
is dead simple to use
doesn’t require me to also solve the problem of where items should ultimately live or how they should be organized long-term.

I just want to save the thing and move on with my life!

Enter JavaScript bookmarklets, which isolate and solve that very first piece of the puzzle: capturing content with one click, and putting it in a reliable spot in a format I can easily access later.

Features

LLM Chat Capture: Dedicated bookmarklets capture conversations from ChatGPT, Claude, and Gemini with one click, formatting them into clean Markdown for the Field Note #0014 A repository for storing vast pools of raw data in its native format. In a personal context, it serves as a low-friction catchment basin where articles, chats, and ideas land before they are structured or organized. View Topic .
Social Media/HTML Article Capture: More dedicated bookmarklets extract core content and metadata from social media (Reddit, LinkedIn) feeds and long-form articles.
Local Sync: This process integrates with a background watcher that moves downloads directly into a specified directory for immediate cataloging.
Local Control: No third-party anything.

How it Works

The bookmarklets trigger a browser download of a timestamped Markdown file containing the content (different bookmarklets handle different types of page/format). A local launchd service (on macOS) watches the Downloads folder and automatically moves these files into the specified data lake directory. Whenever a new file hits the data lake directory, a “catalog” file is automatically updated with the file name and date, and tagged as UNPROCESSED to mark it for the next stage of ingestion.