Built for teams that care about structure, longevity, and true ownership of their data.
.kgl files — diff in git, read with any editor,
parse with the open kgl-lang PyPI package. No proprietary format.
docker compose up and you're running. No account needed,
no credit card, no phone home. Core is MIT-licensed and always will be.
Three concepts. One format connecting all of them.
.kgl files — commit them, back them up.
.kgl files live in MinIO or on disk. Export, migrate, open in a text editor.
No lock-in. No asking permission to leave.
Connect a source, decide what starts it, map its fields — Flow writes the records in as typed nodes.
Most knowledge tools are landlords — your data lives in their database, on their terms, until they raise prices or shut down. Graxeon is different: your knowledge lives in files that belong to you, readable without any special software.
.kgl files with any text editor, any OS
kgl-lang parser on PyPI — write your own tools
@node Attention Is All You Need @type Article @field author: Vaswani et al. @field published: 2017-06-12 @field venue: NeurIPS 2017 @tag ai, transformers, nlp @link Concept/Transformer Architecture @link Person/Ashish Vaswani --- Introduces the Transformer — a model architecture based entirely on attention mechanisms, dispensing with recurrence and convolutions entirely...