#9090th of 102 in Backends, databases and hosting

agensgraph

PostgreSQL fork with a built-in property graph queried via SQL, openCypher and GQL

Stars
1.5k
License
Apache-2.0
Last commit
Oct 2026
Last release
Oct 2026
Language
C
Stars, last 29 weeks+34
21 Apr 2026: 1,485 stars28 Apr 2026: 1,485 stars5 May 2026: 1,486 stars12 May 2026: 1,486 stars19 May 2026: 1,489 stars26 May 2026: 1,489 stars2 Jun 2026: 1,490 stars9 Jun 2026: 1,490 stars16 Jun 2026: 1,493 stars23 Jun 2026: 1,494 stars30 Jun 2026: 1,494 stars7 Jul 2026: 1,495 stars14 Jul 2026: 1,497 stars21 Jul 2026: 1,498 stars28 Jul 2026: 1,498 stars4 Aug 2026: 1,500 stars11 Aug 2026: 1,504 stars18 Aug 2026: 1,505 stars25 Aug 2026: 1,506 stars1 Sept 2026: 1,506 stars8 Sept 2026: 1,506 stars15 Sept 2026: 1,507 stars22 Sept 2026: 1,507 stars29 Sept 2026: 1,509 stars6 Oct 2026: 1,509 stars13 Oct 2026: 1,510 stars20 Oct 2026: 1,512 stars27 Oct 2026: 1,512 stars3 Nov 2026: 1,516 stars10 Nov 2026: 1,519 stars

1,485 on 21 Apr 20261,519 on 10 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

AgensGraph is a fork of PostgreSQL, not an extension, that stores graphs and relational tables in one database and commits them in one transaction. A graph is a schema, a label is a table, and vertices and edges are rows. Queries can mix SQL, openCypher and ISO GQL in a single statement, and an AI section lists MCP servers and LangChain, LlamaIndex, LightRAG and cognee integrations.

Who it is for: Teams wanting a graph database inside PostgreSQL, incl. GraphRAG

Strengths

  • Graph and relational data share one transaction, so joins between Cypher results and tables work in SQL
  • Keeps PostgreSQL tooling: pg_dump, pg_upgrade, row-level security, streaming and logical replication
  • Works with pgvector for vector, full-text or hybrid search followed by graph traversal
  • Variable-length paths plus shortestpath(), allshortestpaths() and dijkstra() built in

Weaknesses

  • Fork of PostgreSQL, so it cannot be added to an existing stock PostgreSQL server as an extension
  • Only a Python driver adds graph types; other languages use plain PostgreSQL drivers
  • Upgrading from 2.17 requires following a dedicated upgrade procedure to 2.18
  • AI integrations ship as separate Python packages, with no non-Python equivalents listed

What it needs

  • no GPU
  • Models: pgvector embeddings, MCP clients, LangChain, LlamaIndex, LightRAG

Quick answers

Is agensgraph open source?

Yes. Its code is on GitHub under the Apache-2.0 license.

Can you run agensgraph with Docker?

Yes. Its repository has a Docker setup: a Dockerfile, a Compose file or a published image.

Also in Backends, databases and hosting

See all 102
Also in Backends, databases and hosting
RankProjectScore
1bunJavaScript and TypeScript runtime with bundler, test runner and package manager96.7k starscustom license93 out of 100
2strapiHeadless CMS that generates REST and GraphQL APIs from content models Live demo ↗ (opens in a new tab)73.5k stars+245 (+0.3%) in 30 dayscustom license92 out of 100
3traefikHTTP reverse proxy and load balancer that auto-discovers routes from orchestrators65.5k stars+446 (+0.7%) in 30 daysMIT92 out of 100
#4appwriteBackend with auth, databases, storage, functions and an MCP server for agents Live demo ↗ (opens in a new tab)58k stars+463 (+0.8%) in 30 daysBSD-3-Clause92 out of 100
#5medusaNode.js commerce platform with modular products, carts, orders and payments37.2k starscustom license92 out of 100