Re:Likes

Re:Likes is an open source JavaScript library that lets readers mark individual passages of text with like or dislike reactions rather than rating an entire piece of content with a single button. It's built for publishers, researchers, and platform developers who want to know which sentences resonate and which fall flat, not just whether the overall article got a thumbs up.

The problem it addresses is the bluntness of traditional reaction systems. A like button at the bottom of an article tells you almost nothing about what worked. Did readers love the introduction but lose interest halfway through? Was one paragraph controversial while the rest landed fine? Re:Likes captures that nuance by tying sentiment to selection. Readers highlight a passage, choose a reaction, and their input persists as a colored overlay on the text.

Under the hood, Re:Likes builds on the Clean Selection library for cross browser text highlighting. When a user selects text and picks a reaction, the library stores that selection locally and optionally syncs it to a backend of your choice. Reactions render as soft overlay clouds that stay visible on the page, and users can revise or erase them later. The result is a heatmap of engagement you can aggregate across readers to see which passages triggered the strongest responses.

Integration requires adding the library to your frontend and configuring rendering options. You control cursor styles, interaction modes, and how overlays appear. It supports both touch and pointer input, so it works on mobile and desktop. Registered users can have their reactions persisted to a backend, while anonymous visitors can still interact with local storage fallback.

Who benefits most are content publishers testing article drafts before wider release, UX researchers studying how people actually read long form content, and community platforms where nuanced feedback on posts matters more than simple upvotes. If you run a comments section, integrating Re:Likes into replies could surface exactly which parts of a discussion people agree or disagree with.

What sets Re:Likes apart is the granularity. Most reaction systems operate at the document level. This one operates at the sentence level, which changes what you can learn from reader behavior. Instead of asking did they like it, you can ask what specifically did they like.

The library is open source and available on GitHub. There's no hosted SaaS version, so you'll need to integrate it into your own stack. That means it's free to use but requires developer effort to deploy and maintain.