Graph shows what a text actually says: the names it uses and the links it states between them.
First step: read the measured result below, then open the full proof in the documentation on GitHub.
Written in Python. Its main work needs no added code libraries.
Measured on 2026-10-15 against a hand labeled set of 20 documents, with 133 gold entities and 66 gold relations: entities F1 0.904 and relations F1 0.598.
The misses are named, not hidden: the tool misses names its word lists do not know, such as Warsaw and Neil Armstrong, and it cannot follow pronouns across sentences.
Full scores, misses, and how they were measured: The proof in the documentation.
You give Graph a piece of text. It finds the people, groups, places, dates, and tools the text names. It also finds links the text states, such as who founded a group, where a group is based, and who works for it. It puts the names and links in a map you can search and check.
It works by fixed rules and word lists. It does not use a large language model. It was built to learn how far plain rules can go on their own.
The layer below: Cohort, belonging
The layer above: Machine, autonomy
The research under this layer: Foundations, Truth and shared knowledge
Built on Cohort: belonging.
Unlocks the ability to discern.