Foxglove is the easiest and fastest way to accelerate and scale your robotics and embodied AI development.
Foxglove 2.x will simply be referred to as Foxglove in this article.
Foxglove is more than a multimodal data visualization tool—it’s a comprehensive platform purpose-built for robotics and embodied AI development. Supporting diverse data formats like MCAP, ROS, Protobuf, and JSON, Foxglove seamlessly facilitates live and recorded data analysis through over 20 visualization panels. It fosters collaboration with shareable layouts and extensible customization, empowering teams to work efficiently. Additionally, Foxglove simplifies uploading, streaming, organizing, and managing petabytes of data. Whether debugging complex 3D interactions, analyzing time-series data, or managing multimodal datasets, Foxglove delivers a modern, flexible, and robust solution designed to meet the challenges of robotics development.
Foxglove empowers over 10,000 robotics and embodied AI developers, while supporting hundreds of robotics companies in accelerating and scaling their development. By streamlining every step of the process—data collection, ingestion, visualization, debugging, and managing vast amounts of robotic data—Foxglove enables teams to collaborate more effectively and innovate faster.
“Foxglove helped us to supercharge our processes. We went from days to minutes when finding the root cause of an issue.” J’aime Laurenson, Product Lead at Wayve
In March 2024, Foxglove announced version 2.0, marking a pivotal shift in the product strategy. This update unified the original open source project, Foxglove Studio, and the Foxglove Data Platform into a single, cohesive product aimed at setting a new standard in robotics observability. Alongside this integration, Foxglove introduced several new features, a new pricing model, and made the decision to discontinue the open-source edition of Foxglove Studio.
The move to close-source Foxglove 2.0 was driven by the desire to provide a more integrated and robust solution for robotics developers, addressing the unique challenges associated with multimodal robotics data. By consolidating offerings, Foxglove enhances developer experience and delivers a comprehensive development and operations platform tailored to the needs of the robotics community. Leading up to, and since, the launch of 2.0, the codebase has undergone significant transformations, including the addition of numerous panels, support for various data sources, improved layout management, increases in performance, and loads of new features.
The last update to Foxglove 1.x (open source) was in February 2024. Since then, there have been over ~2000 commits to Foxglove’s commercial platform (Foxglove 2.x).
Despite this transition, Foxglove remains committed to supporting open formats and continues to maintain and release other open-source projects, most notably MCAP.
“The lift on building custom robotics development tooling, specifically visualization, is so high that it detracts from the goal of building features and autonomy for the robots.” Vibhav Altekar, Co-founder & VP of Software, Saronic
At a high level the differences as of January 2025 are:
Loading identical datasets in the corresponding Plot panels, Foxglove renders data 3 times faster than Foxglove 1.x. Similarly, when working with the same temporal data in the State Transitions panels, Foxglove delivers over 11 times the rendering speed. For developers repeatedly analyzing data throughout their day, these faster load times translate to substantial time savings and a smoother, frustration-free experience due to the enhanced performance.
When viewing compressed video across up to 16 Image panels, Foxglove delivers a fivefold improvement in rendering performance for CompressedVideo topics. This includes significant enhancements to video playback, ensuring a low-latency experience even when visualizing multiple CompressedVideo topics simultaneously.
The Foxglove desktop app on Mac leverages Metal, Apple’s high-performance graphics API. This reduces CPU overhead and provides near-direct access to GPU resources, resulting in substantial performance improvements and a smoother user experience.
Foxglove 1.x struggled with z-fighting issues, which affected the sharpness of its visualizations. To address this, Foxglove introduced a new setting for enabling a logarithmic depth buffer. This enhancement significantly improves rendering precision, particularly when working with large distance scales, ensuring sharper and more accurate visualizations.
The Map panel has been completely overhauled to enhance its rendering efficiency and usability, introducing several exciting new features too:
These are just a few of the many performance enhancements introduced to Foxglove, designed to significantly elevate your daily experience as you work tirelessly to advance your robot’s autonomy.
The Foxglove Image and 3D panel supports additional encodings including over 15 raw image formats and the four leading image and video encodings:
Foxglove supports the playback of multiple local files simultaneously, allowing you to seamlessly view combined data. Simply select Open local file(s), choose multiple files from your system, and immediately begin playback as if it were a single recording.
This update provides the same experience as uploading multiple files to Foxglove data management or merging them with the mcap merge CLI command. For consecutive recordings, you can play through the files from start to finish without interruption. For overlapping files, topics from all recordings will play in log_time order, provided that transforms belong to the same tree and all messages on the same topic share the same schema. This enhancement simplifies working with fragmented or overlapping recordings, streamlining your workflow.
Foxglove goes even further with interactive playback controls too, allowing you to adjust and trim playback time, view timestamps in different formats, and scrub back and forth to specific moments effortlessly. This streamlined approach saves time and ensures an intuitive experience for exploring and analyzing your robotics data.
Foxglove allows you to create shareable links that reference data associated with your devices using their assigned names in Foxglove. To create a link, specify the following parameters:
Foxglove also offers custom extensions that allow you to create unique visualizations tailored to your needs. These extensions can be shared through an extension registry, making it easy for users to find and implement custom solutions.
Organization Layouts enable you to create, edit, and share your custom visualization layouts across your entire team. Additionally, Foxglove provides a central repository where team members can upload, access, and explore shared data. Teams can quickly locate events of interest and stream relevant data directly into visualizations for deeper analysis. This collaborative approach streamlines debugging and fosters more effective teamwork, making Foxglove a valuable tool for scaling your development.
Foxglove streamlines data collection and management with the Foxglove Agent, enabling real-time or on-demand ingestion of recordings. Whether you’re importing data, uploading local files, or connecting to a live robot, the Foxglove Agent ensures seamless data serialization, transportation, and processing.
Foxglove handles asynchronous data streams from diverse sources or massive datasets with ease, indexing data by device, time, and topic for organized management. With customizable retention policies and smooth integration into existing data pipelines, you can optimize storage and maintain efficiency. Flexible access through the UI, API, or CLI lets you retrieve exactly the data you need for streamlined exploration, analysis, and decision-making.
Foxglove provides a central data repository that allows you to upload, explore, set retention policies, and stream your robotics data.
The Timeline view in Foxglove displays each device’s data availability across time. It highlights whether data is ready to stream, actively being processed, or available for import from your robot’s disk using the Foxglove Agent. Use the Timeline to zoom in on specific devices and time ranges to see the status of your data. From there, you can fetch data that’s available for import, visualize it for in-depth analysis, or export it for use with other tools.
Devices in Foxglove serve as representations of all the real and simulated robots your organization tracks. They act as a central reference point for managing and organizing data collected during operations.
When importing data, you can associate each recording with the device that generated it. This linkage ensures that all data is properly categorized and traceable, making it easier to analyze recordings, understand context, and troubleshoot specific robot behaviors. By organizing recordings under their respective devices, Foxglove streamlines data management and enhances collaboration across your team.
Events are key markers within your organization’s recordings, allowing you to quickly identify and categorize specific moments or periods of interest. These events streamline data exploration and help you focus on relevant subsets, improving efficiency during analysis.
Each event is tied to a specific device and requires essential information: a timestamp and a duration. For instantaneous events, the duration is set to zero. Events can also include optional metadata, providing additional context or details to enhance searchability and categorization. By leveraging events, you can simplify the process of locating and analyzing critical moments in your recordings, making your workflows more precise and efficient.
Foxglove continues to enhance its platform with powerful updates to improve usability, flexibility, and data exploration:
Timestamp flexibility: visualization playback supports timestamps in RFC 3339 (ISO 8601), 24-hour, 12-hour, and epoch formats, offering customization for your preferred time display.
Simplified Tab panel renaming: renaming Tab panels is now easier, helping you efficiently organize complex layouts without switching views.
Recent cloud data sources: recently accessed cloud data sources now appear in the “recent sources” list on the dashboard and left-hand menu, marked with a ☁️ icon for easy identification.
Enhanced scatter plots: scatter plots now display X and Y values directly in tooltips and legends as you hover over data points, improving precision during analysis.
Selective recording downloads: choose specific topics to export and trim the time range when downloading a recording. This feature is also available via the /data API and Timeline page for greater control over your data.
Timezone adjustments: quickly change the timezone for data display through the app’s top navigation bar.
Refined playback range updates: adjusting playback time ranges automatically updates the seek time, streamlining data review.
WebSocket protocol update: a new serviceCallFailure operation in the Foxglove WebSocket protocol enables custom panels and the Service Call panel to manage call failures effectively.
Copying logs: a new button in the Log panel allows you to copy logs directly to your clipboard, speeding up debugging and collaboration.
Advanced data filtering: added support for additional operators (!=, <, <=, >, >=) in message path filters. Execute complex queries like /markers/annotations.markers[:]{id<5000}{id>1000}.id for granular data exploration.
Fullscreen panel mode: easily switch any panel to fullscreen by clicking the ⛶ icon in the top-right corner, providing an immersive data view with a single click.
Foxglove continues to prioritize usability and efficiency. While this list highlights some of the key improvements, the journey of enhancing your user experience is ongoing, ensuring that Foxglove remains a powerful, adaptable tool for robotics and embodied AI development.
Foxglove is a team of diverse, highly skilled, and experienced professionals committed to transforming the robotics developer experience. Built on insights from hundreds of customers who push the boundaries of robotics development daily, Foxglove’s platform is designed to deliver an unparalleled developer experience. By leveraging these insights, Foxglove aims to empower developers, streamline innovation, and contribute to increasing the global impact of robotics.
“We expected just easier visualization, but it turned out to be much more—unlocking invaluable efficiency for all our developers.” Robert Sun, Founding Engineer, Dexterity
Check out the Why Foxglove guide to learn more, sign up and get started for free, and join our discord community to share your projects and experiences and work with the best of them!
