About

A European initiative to structure the web.

Company data is locked in Google Maps and expensive data brokers, and much of it is missing or outdated, especially in Europe. SitesGraph reads the public web instead.

What we do

A background crawler follows the links between websites to find new ones, European domains first. Each site is read, analysed and stored as one structured record. The index is searchable, exportable and available through an API. We treat every website as an organisation and the web as one big directory.

Why

Europe wants its own data and search infrastructure, and small language models now make extracting structured data from any website cheap and fast. Longer term, this is a base for a European search engine and for a Global Trends system: following companies over time to learn what works.

Who

SitesGraph is built and operated by Autonomous BV, an AI research lab that studies whether AI can invent, build and run companies. SitesGraph is one of those ventures, and it also feeds the others with data. Contact: contact page.

Under the hood

One Node.js process with an embedded database (RocksDB) and search engine (DuckDB). No paid APIs and no AI in the extraction. Details are in the docs.

Credits and licences

• Technology rules from webappanalyzer, the open-source continuation of Wappalyzer (GPL-3.0).
• Contains data from GeoNames (CC BY 4.0), used to recognise cities.
• Icons by Lucide.