Graffy is a toolkit for graph-centric APIs. It has capabilities comparable to GraphQL and Firebase.

Graffy supports complex, expressive live queries - with multiple levels of resource expansion and pagination - based on a novel application of set theory and CRDTs.

import Graffy from '@graffy/core';
import GraffyClient from '@graffy/client';

const store = new Graffy();
store.use(new GraffyClient('/api'));

const query = {
  posts: [{ last: 10 }, { // Pagination parameters
    title: true,
    author: { name: true }
  }]
};

for await (const state of store.watch(query)) {
  // Iterates each time relevant data changes on the server.
  console.log(state);
}

Graffy provides live queries, which give clients a real-time view of the data they need. Graffy supports complex queries with nested graph traversals and pagination, while exposing a simple and intuitive API for building clients and servers.

Graffy was inspired by (and borrows from) Facebook's GraphQL and Netflix's Falcor. Compared to GraphQL, Graffy offers a more familiar data model, true live queries and more efficient caching. Compared to Falcor, it provides cursor-based pagination and real-time subscriptions.

Unlike GraphQL resolvers and Falcor data providers, Graffy providers can be composed like Express/Koa middleware. This allows authentication, validation, custom caches and resource limiting to be implemented in a straightforward manner.

Graffy providers can also perform efficient bulk reads from underlying data stores (for example by constructing optimized SQL queries). This is particularly hard to do with GraphQL (see dataloader) and Falcor.

The graffy metapackage exports a constructor for a Graffy store in its default configuration.

All the Graffy packages are published under the @graffy scope on NPM.

⌛ = On the roadmap.