Radar / AI robotics / Cerulion
Cerulion: a ROS 2-compatible robotics runtime with deterministic replay
A Rust robotics runtime that runs ROS 2 nodes unmodified alongside native Cerulion nodes, with multi-process graphs, 4.08µs p50 round-trip latency for 1MiB images, and byte-exact deterministic replay from recorded MCAP bags.
Why it matters
Every node free-runs in its own process for low latency, yet the whole graph replays deterministically — record a run, re-execute it, and compare each byte.
What you could build with it
A robotics team could replay a failed warehouse run byte-for-byte to find the exact sensor reading that confused the arm, then fix the controller and prove the fix on the recorded data before touching the real robot.
Does it hold up?
Too early to judge — launched 12 Nov with 115 stars and no independent deployments reported yet; the deterministic-replay claim is the thing to verify on real hardware.
Built with Cerulion
- Show HN: Cerulion – ROS 2-compatible robotics runtime with multi-process replayhackernews · Show HN launch thread where the authors explain the DAG scheduler, per-input bag recording, and the 4µs latency claim.
- Show HN: Cerulion - ROS 2-compatible robotics runtime with multi-process replaybetterhackernews.com · The makers' Show HN post: iceoryx2 shared memory, 4.08us p50 round trips for 1MiB images, and deterministic resimulation with byte-level verification.
- Cerulion docsdocs.cerulion.com · Official docs for the Rust runtime, its DAG scheduler, and the ROS 2 compatibility layer.
Learn more
First spotted on github: source.
More AI robotics
RuView: see people through walls with WiFi, no camerasOpen-source Rust system that turns ordinary WiFi signals into real-time pose estimation, vital-sign monitoring, and…robotics · JEV 0.76Runway Praxis-1: open-weight robot model that learned to act from videoRunway released Praxis-1, a 27B-parameter world-action model, with open weights, inference and fine-tuning code under…robotics · JEV 0.66NVIDIA GR00T 2: a world-action robot model that learns from human videoNVIDIA's open frontier world-action model predicts future video and robot actions together, hitting 71% one-shot…robotics · JEV 0.66RoboParty RP1: a full-stack open-source humanoid robot for embodied AI researchRoboParty's 1.3-metre humanoid, launched 16 October and back in coverage on 2 November, comes in 24-DoF standard and…robotics · JEV 0.62Gemini Robotics 2.5: DeepMind's most capable robot model, tested on a live Mercedes lineGoogle DeepMind introduced Gemini Robotics 2.5, its most capable vision-language-action model, reporting 30-100%…robotics · JEV 0.61T-Rex: a robot manipulation AI that reacts to fingertip touchPanasonic, UC Berkeley, NVIDIA and Stanford built T-Rex (Tactile-Reactive Dexterous Manipulation), a VLA model that…robotics · JEV 0.61
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.