Foxglove SDK​

Python​

• Context is now available for associating channels and sinks.

Rust​

• Some handler types on WebSocket FetchAsset and Service have changed.
• The log! macro properly re-uses matching channels.

C++​

• Context is now available for associating channels and sinks.

Docs: Introduction | Reference: Rust, Python, C++ | Source: github | C++ Artifacts: github

Docs: About | Install | Upgrade

🎨 Set URDF opacity and toggle outlines​

You can now adjust the opacity of URDF models and toggle their outlines in the 3D panel. This gives you more control over how your robot models are displayed.

🔍 Marquee zoom for Plot and State Transitions panels​

The Plot and State Transitions panels now support marquee zooming. Click and drag to select an area to zoom into.

⌨️ Keyboard shortcuts on desktop​

The desktop app now supports keyboard shortcuts for common actions:

Downloads: Linux: deb amd64 arm64, snap amd64 | Windows: Universal | Mac: Universal

Agent​

This release contains a fix for stalled API calls.

Docs: About, Install | Debian Packages: amd64, arm64 | Binary Files: amd64, arm64

🥳 Initial public release of the Foxglove SDK​

We're excited to introduce the Foxglove SDK, a unified toolkit that simplifies integrating your robotics stack with Foxglove. The SDK provides a single, idiomatic API for live visualization, logging, and remote monitoring.

Today, we're releasing the initial version of the Foxglove SDK for C++, Python, and Rust. Built on a shared Rust core, it ensures consistency and reduces feature drift. You can:

• Log structured data to MCAP files for offline analysis
• Stream live data to Foxglove via the WebSocket protocol
• Use high-level logging APIs for common Foxglove schemas like CompressedImage and SceneUpdate
• Define and log custom message types with minimal boilerplate
• Switch between live streaming and file logging using the same instrumentation

Get started​

Get started in a few lines using the SDK to log your data to MCAP files, stream live data directly to the Foxglove app, or both simultaneously:

import foxglove as fg
from foxglove.schemas import Pose, PoseInFrame, Quaternion, Timestamp, Vector3

fg.start_server()

with fg.open_mcap("foo.mcap"):
    fg.log("/example", PoseInFrame(
        timestamp=Timestamp(sec=0),
        frame_id="ego",
        pose=Pose(
            position=Vector3(x=1.0, y=2.0, z=3.0),
            orientation=Quaternion(x=0.0, y=0.0, z=0.0, w=1.0),
        ),
    ))

Key features​

The Foxglove SDK addresses challenges in integrating custom robotics frameworks and tooling:

• Unified Logging Interface: Record to disk or stream live data using a consistent API
• High-Level and Low-Level APIs: Choose between high-level APIs for common schemas or low-level APIs for custom serialization
• Live Visualization Support: Stream live data to Foxglove for real-time visualization
• Cross-Language Consistency: Shared Rust core ensures consistent behavior across Rust, Python, and C++

Docs: Introduction | Reference: Rust, Python, C++ | Source: github

Docs: About | Install | Upgrade

Brand update​

Introducing the new Foxglove brand.

Downloads: Linux: deb amd64 arm64, snap amd64 | Windows: Universal | Mac: Universal

📮 Plot by any timestamp​

You can now generate plots and state transitions using arbitrary timestamp fields from your messages. This gives you more flexibility in how you visualize your time-series data.

👨‍💼Layout Manager​

We've introduced a dedicated page for managing your Foxglove layouts. The new layout manager makes it easier to organize, share, and maintain your visualization configurations.

🧊 Point cloud cube mode​

Point clouds and laser scans now support a cube rendering mode that displays points as cubes. This is useful when your point cloud data semantically represents voxels.

Downloads: Linux: deb amd64 arm64, snap amd64 | Windows: Universal | Mac: Universal

Docs: About | Install | Upgrade

🚢 Performant API for visualizing data across a time range​

The subscribeMessageRange extension API is now generally available. Using this API, Foxglove can load data in small chunks and cache only a subset of serialized data. By doing this, heap memory usage is reduced while still allowing fast re-reads. As a result, panels can display larger datasets without exhausting memory resources.

The subscribeMessageRange API should be used instead of preload: true and context.watch("allFrames") in order to load data across a data source. These previous APIs are now deprecated.

The reference documentation for the extension API is also now available in the core Foxglove documentation.

⏮️ Quick navigation to recent layouts​

The layout menu now shows a list of recently used layouts at the top for quick navigation.

Downloads: Linux: deb amd64 arm64, snap amd64 | Windows: Universal | Mac: Universal