Rerun

The page showcases projects built with Rerun, including live recordings, code, and tutorials across robotics, computer vision, and machine learning.

Rerun helps robotics and Physical AI teams iterate faster: log from any sensor, visualize in the Viewer, query with dataframes, and train with a dataloader tailored to robotic learning — across one recording or many.

This describes the production backend for Rerun’s data layer, featuring cataloging, byte-range indexing, and retrieval to transform object storage into a queryable, streamable foundation.

Rerun is a visualization toolbox for computer vision. The Rerun blog features technical and company background, as well user stories.

Rerun’s data layer simplifies handling multimodal, multi-rate data by converting 5,015 ARKitScenes scans into recordings directly usable by viewers, dataframes, and PyTorch dataloaders without intermediate export steps.

The page evaluates whether ARKitScenes labels suffice for training, then describes a phone-based pseudo-label layer for dense depth and normals, using laser data to identify remaining measurement needs.

Logos, wordmark animation, brand artwork, and usage guidelines for press, partners, and the Rerun community.

Meet Rerun at the Tech Europe x Modal Physical AI Meetup in Munich. Event details, how to get started, and how to reach the team.

Meet Rerun at the Tech Europe x Modal Physical AI Meetup in Zurich. Event details, how to get started, and how to reach the team.

Thanks for coming to Rerun's talk at ROSCon Global 2026 in Toronto. Meetup, ROS 2 and MCAP guides, training tools, and how to reach the team.

The Rerun SDK is free and open-source, while Rerun Hub provides a scalable, single-tenant deployment of the same data layer in users’ regions.

About Rerun Technologies AB, the company building the unified data layer for Physical AI.

Contact Rerun about the product, Rerun Hub, partnerships, technical questions, privacy, or security.

Rerun’s data primitives support building, analyzing, and optimizing data loops, designed for multi-rate and multimodal data from small-scale recordings to large deployments.

The Data Layer for Physical AI

The terms that apply to the Rerun website, documentation, and open-source SDK and viewer.

Rerun raises **$17 million** to develop a data platform for robotics and autonomous systems, addressing challenges in handling **multimodal, time-based data** in Physical AI workflows. The company’s open-source tools and commercial database target developers in the embodied AI sector, aiming to **speed up development** for major tech adopters. Investors and industry players back the platform’s scalability and flexibility.

Mar 2025SuperbCrewNews article

Rerun secures $17 million in seed funding to develop a specialized database and cloud platform designed for physical AI applications. The investment supports expansion of its infrastructure tailored for AI systems interacting with real-world physical environments.

Mar 2025PYMNTSNews article

Rerun, founded by computer vision and robotics experts, secures $17 million to develop a multimodal data infrastructure stack tailored for robotics, autonomous vehicles, drones, and spatial computing. The platform aims to streamline data management for physical AI applications by integrating diverse data modalities. This funding supports the company’s mission to accelerate advancements in robotics and autonomous systems through specialized data solutions.

Mar 2025Yahoo FinanceNews article

Rerun, a Stockholm-based startup, launches an open-source AI platform designed to bridge cloud-based AI with physical environments like robots, drones, and autonomous vehicles. The company secures **$17 million in seed funding** to develop tools that streamline the transition from AI code to real-world deployment, addressing challenges in data mapping and execution.

Mar 2025TechCrunchNews article