# We turn websites into APIs, and keep them working.

> Markdown copy of https://cleanedweb.com/about/ for AI agents and other readers.
> More: https://cleanedweb.com/llms.txt

ABOUT CLEANEDWEB

Give CleanedWeb the address of a public page with the records you need: job posts, listings, products. It builds a REST API that returns them as JSON with the same fields on every run, checks it, and rebuilds it when the site changes. You pay $1 per 1,000 records it returns.

[Build an API](https://cleanedweb.com/start/)  
[Browse the catalog](https://cleanedweb.com/catalog/)

One record from the Y Combinator Startup Jobs API

- **title**: VP of Finance
- **company_name**: Method Financial
- **company_batch_name**: S19
- **location**: New York, NY / San Francisco, CA / Washington D.C. / Austin, TX
- **salary_range**: $280K - $325K
- **url**: ycombinator.com/companies/method-financial/jobs/…

A saved sample from that API’s page in the [catalog](https://cleanedweb.com/catalog/), not a live run.

01 / WHY IT EXISTS

## Scrapers break quietly.

A site renames a class or moves a field, and the scraper keeps running. It just returns fewer records, or empty prices, and nobody notices until the numbers downstream are wrong. Keeping a scraper honest is maintenance that never ends.

CleanedWeb takes on that work. Each API is built from the site, has a written contract for its fields, and is checked after it ships, so a change on the site leads to a repair rather than bad data in your application.

02 / WHAT YOU CAN RELY ON

## How an API stays correct.

### A contract for every API.

Each API has a schema for its fields and inputs, and a definition version. Pin that version in your code, and a request fails with a version conflict if the API changed, instead of quietly returning different data.

[Versions and changes](https://cleanedweb.com/docs/versioning/)

### Checked daily, rebuilt when it breaks.

Catalog APIs are checked once a day. When one keeps failing, it is rebuilt from the site as it is now and swapped in only after it passes the same checks, with the same API ID and URL. A rebuild that would drop or retype a field becomes a separate API instead.

[Handling source changes](https://cleanedweb.com/docs/source-changes/)

### Limits stated up front.

CleanedWeb reads public pages only, never pages behind a login. Every catalog page shows a saved sample and the API’s known limitations, and a run that returns no records costs nothing.

[Browse the catalog](https://cleanedweb.com/catalog/)

03 / GET IN TOUCH

## What are you working on?

Send the site and the records you need, or book a call with Fatih to talk it through.

CleanedWeb is a product of Vaso Ventures LLC, a Delaware limited liability company. For company, support, or privacy questions, [get in touch](https://cleanedweb.com/contact/).

[Talk with Fatih](https://cleanedweb.com/meeting/)  
[Or start with a website URL](https://cleanedweb.com/start/)
