Readlater is a social bookmarking product for readers who want to save the useful parts of what they read, not just keep another pile of tabs open. It is built around a simple loop: capture text when it matters, organize it into collections when useful, and resurface it later in a cleaner, more focused environment.

The product is built around the problem of recall. People often remember that they read something, but not where they found it, what tab it was in, or which exact phrase they need to search for. Readlater gives those fragments a home by keeping saved texts in one place and making them easier to search, revisit, and share.

At the product level, Readlater combines a browser extension, a web app, collections, semantic search, PDF support, and a reader-mode experience. The extension makes the saving flow lightweight: highlight text, right-click, and save it for later. The app then provides a cleaner place to read, organize, and rediscover those saved texts without ads, sidebars, or the noise of the original page.

Collections are a core part of the experience. A saved text can belong to a topic-specific collection, and collections can be private for personal research or public when the user wants to share what they have gathered. This makes Readlater feel less like a passive bookmark dump and more like a small knowledge base that grows with the reader.

The project also explores more advanced workflows around saved reading history. Semantic search helps surface saved texts by meaning rather than exact keyword matches, PDF uploads make research documents part of the same system, and the MCP integration allows saved texts and collections to be created or retrieved from external AI-assisted workflows.

I work across the product, design, and engineering surface, shaping the experience from the browser-saving flow to the app interface and the systems that make saved knowledge easier to find again.