Skip to main content
Back to trending

pgx

PGX — Lightweight Postgres for AI agents. Written in Rust. Thousands of isolated databases in one runtime. 🐘

  • agents
  • ai-infrastructure
  • ai-infrastructure-development
  • ai-infrastructure-software
  • database
  • open-source
  • postgres
  • postgresql
  • rust
  • rust-lang
View on GitHub
Stars
256
Forks
18
+ today
+5
Created
13d

Ranking data as of October 21, 2026 (UTC).

Star History

Today, hour by hour

256 stars
05

Overview

PGX — Lightweight Postgres for AI agents. Written in Rust. Thousands of isolated databases in one runtime. 🐘 It ranks #47 on GitTiger, gaining +5 stars on October 21, 2026 (UTC).

The project is written in Rust and has 18 forks. It was created 13d ago.

AI-generated summary of the repository README — may contain errors.

What it is

PGX is a Rust project described as "lightweight Postgres for AI agents", running thousands of isolated databases in one runtime. The repository is licensed AGPL-3.0 and tagged with postgres, postgresql and AI infrastructure.

Described capabilities

The README preview lists resource-control features and some performance notes: - Memory limits at session, database and runtime level that refuse runaway usage. - Connection limits, so a neighbour running a 50-connection migration does not starve others. - cgroups as a kernel-level last line of defense under a memory limit. - Performance trade-offs the README states against PostgreSQL 18: short statements are about 0.05–0.15 ms slower, and large single-threaded scans are 1.5–2.2× slower.

The preview is truncated, so the full feature list was not supplied.

Getting started

No installation or usage commands were supplied. Read the README at https://github.com/pgrundev/pgx for setup instructions, and review the AGPL-3.0 license terms before use.

Installation
git clone https://github.com/pgrundev/pgx.git
cd pgx
# see README for setup

Frequently Asked Questions

What is PGX meant for?

Per the description, it is a lightweight Postgres for AI agents, hosting thousands of isolated databases in one runtime.

Is it slower than PostgreSQL?

The README preview says short statements are about 0.05–0.15 ms slower than PostgreSQL 18, and large single-threaded scans are 1.5–2.2× slower.

How does it limit resource use?

The README preview mentions memory limits, connection limits and cgroups.

What language and license does it use?

Rust, under AGPL-3.0.