Vuforia

Vuforia
Information
Type	Augmented reality software development kit
Industry	Augmented reality
Developer	PTC (formerly Qualcomm)
Operating System	Android, iOS, Universal Windows Platform
Supported Devices	Smartphones, tablets, HoloLens 2, Magic Leap 2
Release Date	2010 (as QCAR SDK)
Website	https://developer.vuforia.com

Property "Developer" (as page type) with input value "PTC]] (formerly Qualcomm)" contains invalid characters or is incomplete and therefore can cause unexpected results during a query or annotation process.

Vuforia is an augmented reality software development kit (SDK) for creating applications that recognise and track real-world images and objects through a device camera and overlay digital content on them. It uses computer vision to detect planar images, three-dimensional objects, and physical spaces, then computes the camera pose so virtual content stays registered to the real scene. The current product is branded Vuforia Engine and is developed by PTC, an American software company, after originating inside the chip maker Qualcomm.

Qualcomm built the platform to demonstrate the AR capabilities of its mobile chipsets, releasing it for Android around 2010 and for iOS in 2011, and partnering with the Unity game engine so developers could build AR experiences without writing low-level vision code. Qualcomm sold the business to PTC for $65 million in a deal announced on October 12, 2015. At that time Vuforia had powered more than 20,000 apps with over 200 million installs across developers in 130 countries, and was used by 37 of the Interbrand 100 brands. Under PTC the SDK has been positioned for enterprise use in manufacturing, service, and training, while remaining available to general developers.

History

Qualcomm first shipped the technology as the QCAR (Qualcomm Augmented Reality) SDK, releasing version 1.0 in 2011, and rebranded it as Vuforia in 2012. Version 2.0, released on December 20, 2012, added a Cloud Recognition Service that lifted image recognition beyond the roughly 100 targets that could be stored on a device, allowing an app to match against more than one million images held in the cloud. The same release introduced User-Defined Targets, which let an end user pick an image at runtime as a tracking target, and added Unity Play Mode webcam testing.

PTC completed the acquisition by the end of 2015 and folded Vuforia into a broader AR strategy alongside its product lifecycle management and IoT software. On March 29, 2021, PTC announced Vuforia Engine Area Targets, which it described as the first commercial offering to support persistent AR experiences across spaces up to 300,000 square feet, built from 3D scans captured with Matterport, Leica, and NavVis hardware.

How it works

Vuforia Engine extracts natural features from the camera image and compares them against a stored target resource database. When it finds a match it tracks the target frame to frame and reports the pose of the camera relative to it, so an application can render 3D content anchored to the physical target. Target databases can live on the device or, for large catalogues, in PTC's Cloud Recognition service.

For spatial computing features that go beyond a single marker, Vuforia uses a layer called Vuforia Fusion, which builds on the platform's own computer vision together with the underlying device frameworks ARKit, ARCore, and Windows Holographic. Fusion lets Vuforia provide consistent device pose, anchors, and environment understanding across hardware that exposes those native capabilities at different levels.

Target types

Vuforia Engine recognises several categories of target, which determine what kind of physical thing an app can attach content to.

Target type	Description
Image Targets	Flat images such as printed media, packaging, or posters, matched by their natural feature points.
Multi Targets	Several Image Targets arranged in a known geometric configuration, for example the faces of a box.
Cylinder Targets	Images wrapped onto cylindrical or conical surfaces such as bottles and cans.
VuMarks	Customisable markers that double as a brand-styled graphic and an encoded data carrier (for example a serial number or URL), designed as a successor to the conventional matrix barcode for AR.
Model Targets	Real objects recognised from their shape using existing 3D CAD data or a 3D scan; the object must be geometrically rigid.
Ground Plane	Detection of horizontal surfaces such as floors and tables so content can be placed in the environment without a printed marker.
Area Targets	Whole indoor spaces captured as a 3D scan, supporting persistent AR across areas up to 300,000 square feet.

Platform support

Vuforia Engine builds AR applications for Android, iOS, and Universal Windows Platform devices, and integrates with the Unity game engine as a development environment. It also supports optical see-through stereo eyewear: Microsoft's HoloLens 2 remains a primary supported headset, while support for the Magic Leap 2 has been deprecated and the Vuzix M400 is no longer supported. Area Targets are supported only on devices that expose ARKit or ARCore, as well as the Magic Leap 2 and HoloLens 2.

PTC distributes Vuforia Engine releases through a public repository. The latest published version is 11.3.4, dated June 18, 2025, and recent releases have added support for Unity 6 and raised the minimum supported ARCore version. PTC has marked support for HoloLens 2 and the Universal Windows Platform as deprecated, advising that no further fixes are planned and that projects targeting those platforms remain on version 11.4 until support is removed in a future release.

Product family

Beyond the Engine SDK, PTC sells several Vuforia products aimed at enterprise AR: Vuforia Studio, which converts existing 3D CAD and product data into AR experiences; Vuforia Chalk, a remote assistance tool that lets an expert annotate a technician's live camera view; and Vuforia Expert Capture, which records expert procedures into step-by-step AR instructions for frontline workers.