Everything you need to give agents the web.

Read any page as clean text

Hand Refflux a URL and get back tidy, model-ready text - rendered the way a real browser sees it, with the nav, ads, and boilerplate stripped out. One page, one credit.

Walk a whole site in one call

Give Refflux a starting URL and a budget; it follows the links, skips the duplicates, and returns one clean document per page - the fastest way to seed a knowledge base.

Search the web and read the results

Run a query and get ranked results with the page contents already pulled. One call does the finding and the reading, so your agent never has to make two.

Pull typed data, not just text

Describe the shape you want and Refflux returns matching records - prices, listings, contacts - checked against your types, from pages that were never meant to be an API.

Know the company behind a domain

Point Refflux at a domain, email, or name and get a filled-in company profile back - industry codes, a hosted logo, and the real merchant behind a payment line.

Know the moment a page changes

Tell Refflux what to watch and it checks on a schedule, diffs what matters, and posts a signed webhook the instant something you care about moves.