Rerun

Rerun is an open-source data platform designed for robotics and spatial AI. It provides unified infrastructure to ingest, visualize, annotate, query, and transform robotics data from collection to training. The platform consists of two tiers: Rerun Open Source, which offers local visualization and data preparation with support for multi-format log ingestion, data modeling, and dataframe queries; and Rerun Data Platform, a managed cloud solution for large-scale data management, indexing, search, retrieval, and enterprise security. Rerun enables teams to spot interesting moments in robot logs, expand to surrounding context, and experiment with training data pipelines. It is used by organizations like Hugging Face's LeRobot, Google's Brush, NVIDIA's PyCuVSLAM, Ultra, and Meta's Project Aria. The platform features a powerful visualizer that runs natively and in the browser, supports custom layouts, and provides one-click sharing for collaborative debugging.

Key Features

• Multi-format robotics log ingestion
• Unified data model for recording and tabular data
• Dataframe queries optimized for robotics
• Incremental data transformation and schema evolution

Key Benefits

• Visualize and debug robot runs consistently
• Powerful dataframe queries optimized for robotics data
• Multi-format robotics log ingestion
• One-click sharing and collaboration