GPUG: A GPU-accelerated graph visualization engine built with Zed's GPUI, exploring how far Rust and GPUI can take interactive graph rendering.

The origins of graph theory are humble, even frivolous.

~ Norman L. Biggs

---

gpug

GPUG is an initial prototype exploration in building a high-performance, interactive network graph visualization library that leverages the GPU via Zed's gpui library. This approach might allow for the visualization and manipulation of complex graphs, and serves as a foundation for a future Rust library focused on graph interactivity and visualization.

REC2026-08-28-090810-under-10MB.mp4

Created with Create GPUI App, to get started with GPUI visit the GPUI documentation or examples provided:

• gpui
• gpui documentation
• gpui examples
• gpui book
• gpui HN threads
• gpui tutorial
• gpui component library
• gpui music player
• gpui-kit
• first gpui app | sheerluck.dev

Usage

Browser examples

Build all visual examples as browser WASM pages:

python3 web/build.py --setup
python3 -m http.server 8080 --directory target/web-examples

Open http://localhost:8080/. See the browser build guide for prerequisites, individual examples, and static hosting instructions.

0. Ensure Rust is installed ~ Rustup

1. To hack on gpug

   git clone https://github.com/jerlendds/gpug
   cd gpug
   cargo run --example kitchen_sink
   # or to watch:
   # cargo install cargo-watch
   cargo watch -x "run --example kitchen_sink"

2. To build gpug

   cargo build --release

Data, analysis and persistence

GPUG includes a canonical Vortex data model, indexed graph queries and worker coordination through gpug::analysis. DuckDB persistence is opt-in. See the integration guide.

Roadmap

Feature	Description	Completed
Proof-of-concept	Is it possible to render interactive network graphs with GPUI?	true
petgraph or graph backends + more?	Choose a graph backend fit for your usecase	false
Large Graph Handling	Capable of visualizing graphs with thousands of nodes/edges?	true
to be continued	at a later date...	false