Build on it
For writing a tool against a running server, an adapter that brings a new kind of data in, or another implementation of the data type itself.
Libraries
ranke-go
The reference implementation, in Go.
ranke-ts
The TypeScript translation of it, for the browser and for Node.
The server and its clients
ranke-db
The server that serves a Ranke-Graph, with the Ranke-Explorer frontend. Not an app in itself: a platform to build apps on.
ranke-tools
Clients that operate on a graph by talking to the server, and the first use cases built that way. Worth reading before writing your own.
What an implementation follows
Normative specification
The rules an implementation follows, each with a stable id, so a conformance case and a code comment can cite the same handle.
RQL query schema
The machine-readable form of the query language.
Conformance vectors
Claim serializations paired with the ids they are offered under, so an implementation can decide its own conformance.
Changelog
What each release changed in the rules, the papers, the vocabulary, and the vectors.
Writing an adapter
An adapter is an ordinary client: it talks to RankeDB over REST or MCP, and its work is deciding how a source becomes claims — what is a node, what is an edge, and what each claim cites. ranke-git is the worked example, and it is small enough to read in a sitting.
ranke-tools
The adapters that exist, and the shape yours can take.