Project Aria Research Kit Release Notes
These SW Release notes are for the Aria Research Kit (ARK), and are for Meta SW that supports Academic Research Partners using Project Aria glasses. To find out how to become a research partner, go to projectaria.com.
Our Open Source Project Aria Tools SW Release notes are posted to Project Aria Tools GitHub repository, using Tags > Releases.
ARK v1.15 is Now Available!
We are excited to announce the release of Aria Research Kit v1.15 today, containing improvements for both user experience and technical performance. This update adds robust debug sharing in Aria Studio and adds a settings page for customization of machine perception job submission. A new version of the companion app enables users to quickly report issues using the “rage-shake” feature. This release also brings significant improvements to MPS Hand Tracking accuracy.
Key Features
Aria Studio
-
MPS Job Debug Sharing: Users can now easily create and submit MPS job failure reports directly within Aria Studio. Attach screenshots, log files, and grant permission for the Aria Team to provide technical support—all from the new interface. You can also track the history and real-time status of your support cases, streamlining communication and accelerating issue resolution.
-
Settings Page: A new GUI settings page in Aria Studio allows you to customize MPS settings with intuitive, accessible controls. This enhancement makes it simpler to adjust relevant options, improving usability and efficiency when debugging or configuring your environment.
Download and Installation Details: Aria Studio one-click Installation;PyPI installation
Aria Companion App v238
- Rage-shake Bug Reporting: Quickly report abnormal behavior by simply shaking your phone! The new rage-shake feature instantly captures the issue and connects you to Aria Support for faster problem resolution.
Download and Installation Details: Mobile Companion App installation
Improved MPS Hand Tracking
We are releasing MPS Hand Tracking v3.0.0 on October 20th. This version reduces loss of hand tracking and reduces the mean keypoint estimation error on our test datasets, resulting in more reliable and precise hand tracking results. All Aria Research Kit users will be migrated to this new version at the time of release. If anyone has a need to continue using the existing version for dataset consistency, please reach out immediately so we can keep you allowlisted for the current version of hand tracking until you are ready for migration. This is a server-side update that does not require any software installation or updates.
Questions, Comments, Concerns?
If you encounter any problems with Project Aria software please contact our support team through one of the following channels:
-
NEW! "Rage shake" in the companion app to report a bug directly to the team.
-
NEW! Report MPS Job failures directly in Aria Studio.
-
Post to Academic Partners Feedback and Support workplace group - discussion, feedback or user support
-
Post to Project Aria Discord - best for discussion, feedback, or user support
-
Email us directly at AriaOps@meta.com
Thank you for being part of the Project Aria community. We look forward to your feedback and to seeing what you build with these new capabilities!
Aria Mobile App v230 is now available
Dear Academic Partners,
The Aria Mobile App v230 is now available. Visit the ARK SW Downloads page to download the latest version of the app for Android. For iOS, select Update in the TestFlight app. Keeping your app up to date helps ensure the best performance and compatibility.
PROJECT ARIA LATEST OS
The latest Project Aria OS build 4975080.310.70 (released April 1, 2025).
QUESTIONS, CONCERNS, FEEDBACK?
If you encounter any problems using Project Aria support, please contact us. There are several ways for getting user support.
We encourage you to post issues to GitHub.
-
Post to Academic Partners Feedback and Support workplace group - discussion, feedback or user support
-
Post to Project Aria Discord - best for discussion, feedback or user support
-
Email AriaOps@meta.com - for feedback or user support
Note: If you are a Research Partner with access to the Aria Research Kit, there are several other support options available, although GitHub is ideal for bug reports and feature requests.
Aria Mobile App v220 is now available
Dear Academic Partners,
The Aria Mobile App v220 is now available. Visit the ARK SW Downloads page to download the latest version of the app for Android. For iOS, select Update in the TestFlight app. Please update the app to access the following features and bug fixes.
BUG FIXES
- Various bug fixes improving the overall usage.
PROJECT ARIA LATEST OS
The latest Project Aria OS build 4975080.310.70 (released April 1, 2025).
QUESTIONS, CONCERNS, FEEDBACK?
If you encounter any problems using Project Aria support, please contact us. There are several ways for getting user support.
We encourage you to post issues to GitHub.
-
Post to Academic Partners Feedback and Support workplace group - discussion, feedback or user support
-
Post to Project Aria Discord - best for discussion, feedback or user support
-
Email AriaOps@meta.com - for feedback or user support
Note: If you are a Research Partner with access to the Aria Research Kit there are several other support options available, although GitHub is ideal for bug reports and feature requests.
Announcing Aria Research Kit v1.14 with 21 Landmark Hand Tracking in MPS
We are pleased to announce the release of Aria Research Kit v1.14 which now makes fully articulated 21 landmark hand joint positions available with Machine Perception Services Hand Tracking output as a new CSV file called hand_tracking_results.csv. Researchers can utilize this long-awaited feature to produce detailed 3D hand pose annotations useful for robotics research, contextual AI, and anywhere else that hand signals may be important.
In addition to 21 landmarks output, we output a six-degree-of-freedom transform at the wrist location to allow for transformation of the landmarks from the hand frame to the device frame, and normal vectors computed at both the wrist and palm landmarks. The transform and normals form a consistent system and the normals also match the normals we’ve computed on previous versions of wrist and palm output.
We encourage hand tracking users to transition to this new output format, where we will be investing in any future maintenance and bug fixing. We will also continue to provide wrist_and_palm_poses output, which can be loaded with prior versions of projectaria_tools, but will not invest in bug fixes or feature additions for this legacy format.
-
Here's a Notebook and Sample Code (Python & C++) demonstrating how to load and use the landmarks.
-
See Documentation on data format definitions.
Reminder for Aria Desktop App Deprecation: As mentioned in previous release notes, the Aria Desktop App will be fully deprecated on May 30th, 2025, at which point requests to MPS will not be possible and documentation for the tool will be removed. Please migrate over to Aria Studio if you haven’t already.
Update your Software to Access These Improvements
Aria Companion App
Aria Mobile Companion App App v215 is now available for installation. Visit to the ARK SW Downloads and Updates to download the latest version for Android. For iOS, select Update in the TestFlight app.
Project Aria Tools
The latest projectaria-tools 1.6.0 contains tooling necessary to load and visualize 21 landmark hand tracking output. See documentation for download instructions.
Aria Studio v1.1.0
The latest version of Aria Studi, v1.1.0 is required to submit requests to the MPS service for hand tracking annotations with 21 landmark output. See documentation for instructions on how to install Aria Studio. Note that starting with this version, we no longer support Mac Intel with a one-click installer and Mac Intel users must use pip for Aria Studio installation.
Questions, comments, concerns?
If you encounter any bugs or issues, please contact our support team through one of the following channels:
- Post in the Project Aria Discord
- Post to Workplace
- Email us directly at AriaOps@meta.com
Announcement: Aria Research Kit v1.13 is now available
We are excited to announce Aria Research Kit v1.13, featuring color correction and devignetting on videos recorded with the Aria RGB camera and a new ability for Aria to operate as a Wi-Fi hotspot for streaming, eliminating dependency on external networks. Download the latest versions of the Aria Mobile Companion App, the Aria Client SDK, and Project Aria Tools to access these improvements today.
Color correction and devignetting for the Aria RGB camera
You may have noticed blueish color distortion in Aria RGB recordings. We’ve now fixed these with consistent gamma curves and standard color temperatures in ARK v1.13. With this update, Aria devices now produce images with more consistent colors, calibrated to a standard color temperature of 5000K. Additionally, a new devignetting API is available in Project Aria Tools to further enhance image quality and accuracy.
Wi-Fi hotspot streaming
A new Wi-Fi hotspot streaming feature allows users to set up a dedicated hotspot using the Aria device, enabling direct and stable streaming without relying on external network infrastructure. Users can enable this feature through the Aria Client SDK or directly within the Aria Companion App.
Update your software to access these improvements
Aria Companion App
Aria Mobile Companion App App v210 is now available for download. Visit ARK SW Downloads and Updates to download the latest version for Android. For iOS, select Update in the TestFlight app.
Project Aria latest OS
Project Aria OS build 4975080.310.70 (released April 1, 2025) contains the changes necessary for RGB color correction and devignetting. Update to this build or newer using the Aria Companion App.
Aria Client SDK
The Aria Client SDK now supports streaming in hotspot mode, diagnostic logging collections for users, support for Python 3.12, RGB color correction, and more. See the ARK download documentation for download instructions.
Project Aria Tools
The latest projectaria-tools 1.5.8 contains the tooling necessary to support RGB camera color correction and devignetting. See the data utilities overview for download instructions.
Questions, comments, concerns?
If you encounter any bugs or issues, please contact our support team through one of the following channels:
- Post in the Project Aria Discord
- Post to Workplace
- Email us directly at AriaOps@meta.com
Aria Research Kit v1.12
We are excited to announce updates to the Aria Research Kit with v1.12. Download the latest versions and access new features for Aria Studio, Project Aria Tools, and the Companion App.
Mobile Companion App
Aria Mobile Companion App v200 is now available for download. Visit ARK SW Downloads and Updates to download the latest version for Android. For iOS, select Update in the TestFlight app.
Bug fixes
Version 200 includes minor bug fixes and refinements to the existing functionality and user experience.
Latest Project Aria OS
Version 200 maintains the latest Project Aria OS build: 4971771.1550.70 from October 3, 2024.
Aria Studio
Aria Studio v1.0.3 includes minor bug fixes and refinements. It is now available via one-click installation and pip-install from PyPi in Mac ARM, Mac Intel, Windows, and Linux.
Project Aria Tools
We are excited to announce Project Aria Tools v1.5.7! Download or update your installation of Project Aria Tools to get access to the latest updates.
See our GitHub release notes to review all Project Aria Tools updates.
New Advanced Image Utility: Image devignetting
Devignetting is a digital image processing technique used to correct uneven lighting in images, particularly those captured by cameras with wide-angle lenses or those that have inherent optical vignetting. Vignetting refers to the gradual decrease in brightness toward the edges of an image, often resulting in a darker periphery and a brighter center.
We now provide official guidance on how to apply devignetting with Python and C++. See the notebook and documentation for more information.
Questions, comments, concerns?
If you encounter any bugs or issues, please contact our support team through one of the following channels:
- Post in the Project Aria Discord
- Post to Workplace
- Email us directly at AriaOps@meta.com
Aria Companion App v195 now Available!
December 13, 2024
Aria Mobile App v195 is now available for download. For Android, visit the ARK SW Downloads and Updates page and download the latest version. For iOS, select Update in the TestFlight app. Please update the app to access the following features and bug fixes:
New & Updated Features
-
A new updates tab dedicated to feature announcements and program updates is now available in-app
-
Our eye tracking calibration feature is now fully compatible with iPhone 16
Bug Fixes
Some bug fixes to improve the overall usage.
Project Aria Latest OS
The latest Project Aria OS build 4971771.1550.70 (October 3, 2024).
Aria Studio - Please Update to v1.0.2 for Bug Fixes
After our v1.0.0/1.0.1 release of Aria Studio on Monday Dec 9th, we identified a bug for users who downloaded Aria Studio with the Windows 1-click installer that caused MPS jobs to fail.
The team worked quickly to resolve the bug, and the fix is now available as part of a new release: Aria Studio 1.0.2. Please update to the latest version (1.0.2) at your earliest convenience.
Questions, Comments, or Feedback?
If you encounter any problems with Project Aria software, please contact our support team through one of the following channels:
-
Post to Project Aria Tools issues on GitHub - preferred
-
Post to Academic Partners Feedback and Support workplace group - discussion, feedback or user support
-
Post to Project Aria Discord - best for discussion, feedback or user support
-
Email us directly at AriaOps@meta.com
Announcing Aria Research Kit v1.11!
December 9, 2024
Aria Studio v1.0.1: Now on Windows, plus some other nice features
Aria Studio is a modern desktop application with a browser-based UI that helps you manage recordings from your Aria glasses, access Machine Perception Services, and visualize egocentric data. We released a beta version of Aria Studio at CVPR this year with support for Mac and Linux. Today, we are releasing Aria Studio v1.0.1, which extends our OS support to include Windows. Aria Studio is now available as a downloadable 1-click installer package in addition to the previously available pip install method. Another new feature includes the ability to submit feedback to the Aria team directly in the application and a section where the Aria team will publish news program updates and important information in the home tab.
Updates for the MPS CLI and the Mobile Companion App
This release also brings the Aria Machine Perception Services Command Line Interface (MPS CLI) to Windows. Download or update your installation of Project Aria Tools on your Windows machine to use the MPS CLI to submit your recordings to Aria Machine Perception Services. Coming soon, we will release v195 of the Aria Companion App on mobile, featuring a new section on the home tab where you can stay up-to-date with the latest news and updates from the Aria team.
Deprecation Plan for Aria Desktop App
For those still using the Aria Desktop App: Today, we are announcing a deprecation plan for the Aria Desktop App. We are no longer onboarding new users to the Desktop App and will shut off the ability to request MPS from the Desktop App on May 30th, 2025. Now with support for Windows, Aria Studio fully replaces the functionality of the Desktop App. Please switch over to Aria Studio as soon as possible and reach out to ariaops@meta.com if you have any questions about the transition.
Feedback and Support
If you encounter any problems using ARK v1.11, please contact us. There are several ways for getting user support.
-
Post to Academic Partners Feedback and Support workplace group
-
Post to Project Aria Discord
-
Email AriaOps@meta.com
You can always request a meeting through our ops team if you need more in-depth discussions requiring higher bandwidth.
Aria Mobile App v185 is now available
October 25, 2024
Dear Academic Partners,
The Aria Mobile App v185 is now available. Go to the Downloads page to get the latest version for Android. For iOS, select Update in the TestFlight app. Please Update the app to access the following features and bug fixes.
BUG FIXES
Various bug fixes to improve the overall usage.
PROJECT ARIA LATEST OS
The latest Project Aria OS build 4970205.890.70 (June 28, 2024).
QUESTIONS, CONCERNS, FEEDBACK?
If you encounter any problems using Project Aria support, please contact us. There are several ways for getting user support.
We encourage you to post issues to GitHub, so that they can be tracked.
If you are a Research Partner with access to the Aria Research Kit there are several other support options available, although GitHub is ideal for bug reports and feature requests.
- Post to Academic Partners Feedback and Support workplace group - discussion, feedback or user support
- Post to Project Aria Discord - best for discussion, feedback or user support
- Email AriaOps@meta.com - for feedback or user support
Announcing Aria Research Kit v1.10
October 9, 2024
We're announcing the release of Aria Research Kit v1.10. This release adds versioning to all of the Machine Perception Services and includes a major update to make the SLAM service more accurate for downstream tasks that utilize 3D results projected into the RGB camera frame. We are also making it much easier for you to visualize, edit, and prepare your Aria recordings for downstream use such as dataset preparation and model training with PyTorch.
MPS Versioning
Starting now, we are introducing the concept of versioning to machine perception services. When we make changes to our SLAM, hand tracking, or eye tracking services, you will have information about which version of the service was utilized for your result computation. This will be available directly at the very end of the summary.json file. Versions will have a major, minor, and bugfix number (e.g. SLAM v1.1.0). We will keep our MPS Service Versions page updated as we roll out new versions so that you are aware of the changes we make.
RGB Online Calibration
Aria excels at enabling computer vision from a 3D perspective. Our SLAM Machine Perception Service provides Aria trajectory and a semi-dense point cloud output that allow computations in a 3D reference frame. Tasks like object detection can be performed in 3D and then be re-projected back into the 2D RGB images. Available as of v1.1.0, we’ve implemented extremely accurate RGB online calibration for all of your downstream tasks that utilize 3D results reprojected into the RGB camera frame. Read the MPS Output Calibration page for more details on how to make use of online calibration.
New Tools for Working with VRS
We know that it takes manual work to prepare Aria recordings for downstream use, so we’ve released some new tools to make things easier. We now provide a lightweight VRS Player App that allows you to easily play the visual and audio sensor streams of a recording from a file menu or by using a simple command. We’ve also released a VRS Command Line Tool that enables easy inspection and editing of VRS files. For example, you can time crop a recording to eliminate an unwanted segment or remove sensor streams that you no longer need in your final dataset after processing through Machine Perception Services. Install both the lightweight VRS Player and the VRS Command Line Tool with the VRS Tools Installer.
Aria Training and Evaluation Toolkit
Preparing Aria recordings for training in PyTorch just got easier with Aria Training and Evaluation Kit (ATEK). ATEK offers an efficient solution to convert VRS files to PyTorch compatible formats, simplifying integration and use within deep learning projects. We’ve also pre-processed two popular Aria datasets with ATEK for direct usage in a data store. And lastly, we’ve created standardized evaluation libraries for 3D object detection and surface reconstruction. Read an in-depth blog post on ATEK for more information and see the ATEK GitHub page to install ATEK today.
Project Aria Now on Huggingface
Finally, we introduce the Project Aria Huggingface Space, which allows researchers to track their model performance, helping measure, compare, and accelerate AI and ML research within the open community. AI models and datasets, released in the Project Aria Hugging Face space, will be supported by ATEK. Where possible, AI models trained using Project Aria data will also be available with full weights and scripts, making it easy for researchers to discover and benchmark existing models against their own methods.
Aria Studio Updates
We've also made some updates to Aria Studio. We now display a version number and have a mechanism for surfacing updates and other key information that we want to communicate to our users. Today, Aria Studio is installed via pip install, and in doing so you'll automatically get these new features. If you haven't yet tried it, please install Aria Studio today.
Any Questions?
If you have any questions regarding this new functionality in Aria Research Kit v1.10 or any issues accessing any of the features, please reach out to our support team at AriaOps@meta.com.
Aria Mobile App v180 is now available
September 17, 2024
Dear Academic Partners,
The Aria Mobile App v180 for Android is now available. Go to the Downloads page to get the latest version for Android. For iOS, select Update in the TestFlight app. Please update the app.
BUG FIXES
Various bug fixes to improve the overall usage.
PROJECT ARIA LATEST OS
The latest Project Aria OS build 4970205.890.70 (June 28, 2024).
QUESTIONS, CONCERNS, FEEDBACK?
Contact Aria User Support by posting here in Academic Partners Feedback and Support workplace group or emailing AriaOps@meta.com.
Additional resources are available at Project Aria Github.