rerun

# Rerun — The data layer for physical AI Rerun is a visualization and logging SDK purpose-built for multimodal, multi-rate data from robotics, simulation, and computer vision pipelines. It captures images, point clouds, transforms, time series, joint states, and video from many sources — robot logs, human-data rigs, simulators, web video — and stores them in shared columnar storage. The built-in viewer renders everything in sync in real time: scrub episodes, compare sensors side-by-side, watch CV pipelines run live. The same data is queryable with dataframes or SQL and streams directly into training, with no export jobs or stale copies. Built in Rust. SDKs available in Python, Rust, and C++. ## What it does Rerun ingests multi-rate, multimodal data from many sources and formats (robot logs, human-data rigs, sim, web video; MCAP, .rrd, LeRobot). The built-in viewer renders everything in sync in real time. The same data is queryable with [dataframes](https://rerun.io/docs/howto/query-and-transform/get-data-out) or SQL, and streams directly into training. Built in Rust on column-chunk storage purpose-built for multi-rate physical data. ## Quickstart ```bash pip install rerun-sdk ``` ```python import rerun as rr rr.init("rerun_example") rr.spawn() rr.log("points", rr.Points3D(positions, colors=colors)) ``` ## Use cases - Ingest robot logs, egocentric/UMI rigs, sim, and web video into one substrate - Run CV pipelines (SLAM, hand tracking, motion retargeting) as table edits - Query raw, intermediate, and derived data with dataframes or SQL - Visualize multi-rate, multimodal sequences across the pipeline - Stream dataset mixes directly to training — no export jobs, no stale copies ## Data types Multi-rate, multimodal, spatial: images, point clouds, time series, tensors, transforms, joint states, video. Preserved end-to-end. ## Getting started - [Python](https://www.rerun.io/docs/getting-started/data-in/python): `pip install rerun-sdk` - [Rust](https://www.rerun.io/docs/getting-started/data-in/rust): `cargo add rerun` - [C++](https://www.rerun.io/docs/getting-started/data-in/cpp) Install the Rerun Viewer binary with `pip install rerun-sdk` or `cargo install rerun-cli --locked --features nasm`. The Python SDK bundles the viewer; C++ and Rust rely on a separate install. ## Documentation - [High-level docs](https://rerun.io/docs) - [Loggable Types](https://www.rerun.io/docs/reference/types) - [Examples](https://rerun.io/examples) - [Python API docs](https://ref.rerun.io/docs/python) - [Rust API docs](https://docs.rs/rerun/) - [C++ API docs](https://ref.rerun.io/docs/cpp) - [Troubleshooting](https://www.rerun.io/docs/overview/installing-rerun/troubleshooting) ## Agent skills This repo ships agent skills for coding agents. Install with: ```sh npx skills add rerun-io/rerun ``` Skills live in `./skills`. ## Status Active development. API is evolving — expect breaking changes. Known limitations: viewer performance degrades with very many entities, and multi-million point clouds can be slow. ## Business model Open-core. Everything in this repository stays open source and free (MIT + Apache-2.0 dual-licensed). Rerun Hub, a scalable catalog for robotic data, is a commercial product for teams building robotics and CV products. ## Citation ```bibtex @software{RerunSDK, title = {Rerun: A Visualization SDK for Multimodal Data}, author = {{Rerun Development Team}}, url = {https://www.rerun.io}, year = {2024} } ```