FlightRadar24 for space – but fully open‑source and physics‑grade accurate. OpenAstroViz streams, propagates, and visualises every tracked object in Earth orbit – satellites, spent stages, debris – in real time.
Mission statement: Democratise space‑situational awareness. Provide a transparent, vendor‑neutral SSA stack that anyone can run locally or fork for research.
- Live orbital map – WebGL globe + timeline scrubber (Sputnik → +1 year forecast).
- CUDA compute backend (v0) – GPU‑accelerated SGP4 & conjunction kernels (RTX/GeForce/Quadro). We start with NVIDIA CUDA for maximum throughput and mature tooling.
- WebGPU compute backend (v2) – Pure‑browser, cross‑vendor path that eliminates native daemons. Targeted once Safari/WebGPU FP32 stabilises.
- Real‑time conjunction alerts – cell‑grid pruning + analytic miss distance solver.
- Space‑health dashboard – daily / weekly / yearly “Top‑5” congestion indicators.
- Story mode – replay Iridium‑Cosmos, Starlink shell deployment, etc.
# 1 – clone & bootstrap
$ git clone https://github.com/openastroviz/openastroviz.git && cd openastroviz
$ ./scripts/setup.sh # installs Rust, CUDA, Node.js + pre‑commit hooks
$ rustup override set stable # ensure the stable toolchain locally
# 2 – build CUDA backend (requires NVIDIA + nvcc)
$ cargo run -p openastrovizd -- bench cuda # benchmarks, STK vector tests
# 2b – run reference Vallado SGP4 sample (CPU)
$ cargo run -p openastroviz-core --example propagate
# 3 – run the web client (connects to local daemon)
$ yarn --cwd web install && yarn --cwd web dev # http://localhost:5173If you don’t have an NVIDIA GPU, skip cuda/ and work on the cpu‑simd reference path or UI issues labelled good first issue.
The validated Vallado SGP4 implementation now lives in the core/ crate and is exercised by regression tests against published state vectors. The earlier Keplerian proof‑of‑concept has been quarantined to docs/archive/poc_sgp.cpp for historical reference.
Standard fork → branch → PR flow. clang‑format / rustfmt / eslint are enforced via pre‑commit. CUDA code follows the style guide in docs/cuda_style.md.
- Code: MIT © 2025 OpenAstroViz Contributors
- Data: CC0 (derived from U.S. Government public‑domain TLEs)