identity.txt

AI agents scrape and summarise the web without context. Companies have no way to tell these agents who they are, what they do, or how they should be described. We needed a standard file format that any organisation could adopt.

Simple JSON spec inspired by robots.txt. Organisations host identity.txt at their root domain with structured fields: name, description, industry, key products, notable facts. AI agents can fetch and parse it before summarising. Open-sourced the spec and reference implementation.

Advocacy project. We’re using it on our own site and encouraging clients to adopt it. The goal is ecosystem adoption so that AI search engines and agents check for identity.txt by default. This is about creating standards, not products.