The web is still full of good writing.

It is simply harder to encounter by accident. Search expects you to know what you want. Social feeds rank what produces a reaction. Recommendation algorithms learn to give you more of what already worked.

Stumble Reads does something smaller: it keeps a reviewed shelf of writing worth meeting and hands you one piece at a time.

The project is independently curated by Marc Vallverdú. Selections are not paid placements. Where the curator has a direct affiliation with an author or publication, that relationship must be disclosed or the piece stays out of the active catalogue.

Stumble Reads publishes its own recommendation notes and links to original publishers. Optional reader mode exists for accessibility, but the original source is always the primary destination.

Privacy-friendly Umami analytics, served by dashboard.remarcablevc.com, records page views and aggregate actions such as stumbling, opening an original source, saving, or sharing. Event properties can include a recommendation title, publication, category, save state, or sharing method; there are no reader accounts or cross-site advertising profiles. Do Not Track is honored, and adding ?qa=1 disables analytics on that device. Viewed and saved recommendation slugs remain on the reader’s device.