Dgraph

Dgraph, developed by Dgraph Labs, Inc., is an open-source distributed graph database designed for real-time use cases at terabyte scale. It is implemented as a sharded, distributed system that processes queries in parallel and provides distributed ACID transactions, consistent replication, linearizable reads, snapshot isolation and multi-version concurrency control. Users can operate it either as a property graph database for direct database operations and administration, or generate a GraphQL API and graph backend from a GraphQL schema for application development.

The system is written in Go and released under the Apache 2.0 license. It uses BadgerDB, a Go key-value store also built by the same organization, as its on-disk storage interface for posting lists, with log-structured storage organization, write-ahead logging, and configurable Snappy or Zstandard compression. Queries use DQL, a GraphQL-inspired language, and are served over gRPC and HTTP with JSON, RDF and Protocol Buffers responses; the cluster uses the Raft consensus algorithm in a shared-nothing architecture. Native capabilities include full-text search, regular expressions, geo search, inverted indexes and persistent queries. Official platform support covers Linux/amd64 and Linux/arm64, with Docker recommended for testing and deployment.

Beyond the core database, the dgraph-io organization maintains official client libraries for Go, Java, JavaScript and Python, the Ratel data visualizer and cluster manager, Helm charts, benchmarking tools, and a Docusaurus-based documentation site published at docs.dgraph.io.

Founding story

According to Database of Databases, Manish Rai Jain created Dgraph in July 2015 based on his prior experience at Google, where he led a project to unite data structures for serving web search with a backend graph system. The first version, v0.1, was released in December 2015, with the initial open source release cited elsewhere as 2016. The 2022 funding announcement notes the founders' background in distributed systems and graph systems at scale and their aim to build a graph database platform that could serve as an entire application backend rather than only a database.

Business model

Commercial open source: the database is distributed as an Apache 2.0 licensed community edition alongside an enterprise offering with additional enterprise-ready capabilities, plus a hosted cloud service launched in 2020.

Sources describe an enterprise offering with additional enterprise-ready capabilities beyond the Apache 2.0 community edition, and a hosted Dgraph Cloud service; specific pricing is not disclosed in the sources.

Traction

The main dgraph repository has roughly 21.8k GitHub stars and 1.6k forks with over 6,500 commits, and the organization hosts 109 public repositories including BadgerDB (~15.8k stars) and Ristretto (~7k stars). At the time of the 2022 round the project reported over 18,000 GitHub stars and more than 10 million Docker Hub pulls, and the README states Dgraph is production-ready at version v25 and used in production at multiple Fortune 500 companies.

Latest developments

Following the October 2025 acquisition of the combined Hypermode/Dgraph business by Istari Digital, the dgraph-io GitHub organization states the project continues under Istari Digital's stewardship with a focus on maintenance for the community. Repository activity continues into 2026, with commits to dgraph, badger, dgraph4j, docs and new tooling such as the Argus benchmarking harness and the dgdao Go client/DAO. Documentation is versioned for releases v25.2 through v25.4.