← Back to AR History Series

Metaio SDK — The AR Platform Apple Killed and Turned Into ARKit

Prabhu Kumar Dasari — Senior Unity XR Developer
Prabhu Kumar Dasari
Senior Unity XR / VR / AR Developer · 13+ Years
Tested Metaio SDK including 3D object detection · UAE & India · GITEX Dubai 2024
In May 2015, Apple quietly acquired a German company called Metaio. There was no press conference, no announcement on Apple's website, no transition plan for the thousands of developers who had built products on Metaio's SDK. The SDK simply stopped being available. Two years later, ARKit launched — with capabilities that bore a striking resemblance to what Metaio had been doing. As a developer who tested Metaio including its 3D object detection, I watched one of the most technically impressive AR platforms of its era disappear overnight.

What Metaio Was

Metaio was a German augmented reality company founded in 2003, originally as a spin-off from Volkswagen's research division. It began building AR tools for industrial applications — overlaying assembly instructions onto car parts during manufacturing — before expanding into mobile AR development as smartphones became capable enough to run computer vision in real time.

By 2013 and 2014, Metaio had built what most AR developers considered the most technically advanced mobile AR SDK available. While competitors like Vuforia dominated the enterprise market through reliability and ease of use, Metaio competed on raw capability — offering tracking features that no other mobile SDK could match at the time.

📌 Developer Context

I tested Metaio's SDK during this period — including its image recognition capabilities and notably its 3D object detection, which allowed the camera to recognise and track real physical objects like bottles or products without any printed marker. This was the feature that set Metaio apart from every other mobile AR SDK of the era.

What Made Metaio Different — The Features That Were Ahead of Their Time

📦
Most Capable on Mobile
3D Object Detection
Metaio could recognise and track real physical 3D objects — a bottle, a product, a piece of machinery — without any printed marker or QR code. The camera identified the object's shape and texture in 3D space and anchored AR content to it. No other mobile AR SDK of the era matched Metaio's reliability on this feature.
🗺️
Early Mobile SLAM
Markerless SLAM Tracking
Simultaneous Localisation and Mapping — tracking the camera's position in an unknown environment without any marker at all. Metaio brought SLAM to mobile AR before it became a standard feature. This is the same fundamental technology that powers ARKit's world tracking today.
😶
Pre-Snapchat
Face Tracking
Metaio included face tracking capabilities — detecting facial features and overlaying AR content on them — years before Snapchat's face filters made this mainstream. The quality was professional-grade, not a gimmick.
🎯
Image Tracking
Image Target Tracking
Like Vuforia, Metaio supported image target tracking — recognising flat printed images and overlaying AR content. I tested this alongside Vuforia during this period. Both were capable, but Metaio's edge was the additional 3D features that Vuforia did not offer.
📍
Location AR
GPS-Based AR
Metaio supported location-based AR — overlaying content at specific GPS coordinates in the real world. This was used for navigation, tourism, and outdoor AR experiences before dedicated platforms like Niantic made location AR mainstream.
🏭
Enterprise Origins
Industrial AR
Metaio's roots were in industrial AR — overlaying assembly instructions, maintenance guides, and technical information onto physical machinery. This enterprise DNA shaped how the SDK approached precision, reliability, and real-world robustness.

3D Object Detection — What It Actually Meant in Practice

The most significant technical differentiator between Metaio and its competitors was 3D object detection. To understand why this mattered, it helps to understand what the alternative was.

With Vuforia and most other AR SDKs of the era, the trigger for AR content had to be a flat, printed image — a carefully designed image target that the SDK had been trained to recognise. This worked well for marketing materials, packaging, and printed media, but it had an obvious limitation: the trigger had to be printed and placed in advance. You could not point a camera at an arbitrary physical object and have it recognised.

Metaio's 3D object detection changed this. By scanning a real physical object — a bottle, a product, a piece of equipment — the SDK could build a 3D model of that object and use it as a tracking target. Point the camera at the same object later and Metaio would recognise it, establish its pose in 3D space, and anchor AR content to it.

🧪 What I Tested — 3D Object Detection

During my testing of Metaio's SDK, the 3D object detection was the feature that genuinely impressed me. Scanning a physical object — something like a bottle — and then having the camera recognise and track that same object in real time felt significantly more advanced than anything Vuforia was offering at the time. The use cases this opened were different in kind from image target AR: you could build AR experiences around actual products, not just their packaging.

Why 3D Object Tracking Was Hard

Detecting and tracking a 3D object in real time on a mobile CPU was genuinely difficult. Unlike a flat image target — which has a fixed, known appearance — a 3D object looks different from every angle, in every lighting condition, and with every partial occlusion. The algorithm had to match what the camera was seeing against a 3D model of the object across all these variations simultaneously, fast enough to update at video frame rates without draining the battery in minutes.

Metaio's implementation was not perfect — it worked best with objects that had distinctive surface texture and sufficient geometric complexity. Highly reflective surfaces, objects with too little texture, or objects with repetitive patterns were harder to track reliably. But for the right use cases, it was the most capable 3D object tracking available on a smartphone at that time.

Metaio vs Vuforia — The Real Comparison

Feature Metaio Vuforia
Image target tracking ✓ Strong ✓ Strong (market leader)
3D object detection ✓ Most capable on mobile Limited / weaker
Markerless SLAM ✓ Yes Not at this time
Face tracking ✓ Yes No
GPS-based AR ✓ Yes No
Unity integration Good ✓ Excellent
Documentation quality Moderate ✓ Better
Enterprise adoption Growing ✓ Dominant
Still available today ✗ Acquired & shut down ✓ Yes (PTC)

The Acquisition — What Actually Happened in 2015

May 2015 — Acquisition
Apple Quietly Acquires Metaio
Apple confirmed the acquisition with its standard response: "Apple buys smaller technology companies from time to time, and we generally do not discuss our purpose or plans." No price was disclosed. No transition plan was announced. No timeline was given to developers still using the SDK.
May 2015 — Immediate Effect
SDK Downloads Disabled, Developer Portal Closed
Within days of the acquisition announcement, Metaio's developer portal went offline. New SDK downloads were disabled. Developers who had active projects built on Metaio were left with existing installations but no support, no updates, and no path forward. There was no migration guide, no successor product announcement, no grace period.
June 2017 — ARKit Launch
Apple Announces ARKit at WWDC
Two years after acquiring Metaio, Apple announced ARKit at WWDC 2017. The capabilities — plane detection, camera pose estimation, motion tracking, lighting estimation — aligned closely with what Metaio had been developing. Apple had spent two years integrating Metaio's technology directly into iOS, making it available for free to every developer targeting iPhone and iPad.

What Developers Had to Do After the Shutdown

For developers like me who had tested or built with Metaio, the shutdown created an immediate practical problem. Any project using Metaio's SDK could continue running on existing installations but could not be updated, submitted to app stores with newer API requirements, or supported long-term. The choices were stark: rebuild using Vuforia, wait to see what Apple announced, or migrate to one of the smaller remaining SDKs.

Most enterprise clients who had Metaio-based applications already deployed chose to run them until they broke rather than pay for an immediate rebuild. The applications that depended most heavily on Metaio's unique features — particularly 3D object detection — had the hardest migration path, since Vuforia's 3D object capabilities were significantly weaker at the time.

⚠️ The Lesson for Developers

The Metaio shutdown was one of the starkest examples in AR history of the risk of building on a third-party SDK without a migration plan. A technology acquisition can eliminate a platform overnight with no warning and no compensation for developers who invested significant time building on it. This risk has not disappeared — it applies to any third-party SDK or platform today.

Metaio's Legacy — Inside Every iPhone

The most significant thing about Metaio is not that it was shut down — it is what happened to its technology afterward. ARKit, released in 2017, made the kind of capabilities Metaio had been building available for free on every iPhone. World tracking, plane detection, image recognition, object detection — these features, which required a specialised commercial SDK in 2014, became standard OS features three years later.

In that sense, Metaio did not really die. It dissolved into iOS. Every ARKit application built since 2017 — every AR game, every retail try-on experience, every industrial AR tool — runs on technology that traces directly back to what a team of German engineers was building before Apple acquired them.

Metaio SDK 3D Object Detection SLAM Tracking Image Recognition Face Tracking Unity Integration ARKit (successor)
💬 Developer Reflection — Prabhu Kumar Dasari, 13+ Years in XR

Testing Metaio's 3D object detection in those years gave me an early sense of where AR was heading — toward a world where any physical object could become an AR trigger, not just carefully designed printed images. The shutdown was frustrating precisely because the technology was genuinely impressive and the developer community had invested real time in it. But in retrospect, Metaio's acquisition was the best possible outcome for the technology itself. It went from a commercial SDK that relatively few developers used to the foundational AR framework on over a billion devices. The ideas survived even though the company did not.

Frequently Asked Questions

Why did Apple acquire Metaio?

Apple never officially disclosed its reasons for acquiring Metaio. The widely accepted view in the AR developer community is that Apple wanted Metaio's core computer vision and AR tracking technology — particularly its SLAM implementation and 3D tracking capabilities — to build ARKit. The acquisition gave Apple a significant head start in mobile AR development, allowing ARKit to launch in 2017 with mature, battle-tested tracking technology rather than starting from scratch.

Is ARKit based on Metaio technology?

Apple has never officially confirmed the technical relationship between Metaio and ARKit. However, the timeline — Metaio acquired in May 2015, ARKit launched June 2017 — and the capability overlap between Metaio's advanced features and ARKit's launch capabilities lead most AR developers and analysts to believe Metaio's technology was central to ARKit's development. The two-year gap between acquisition and ARKit launch aligns with the time needed to integrate and productise an acquired technology at Apple's quality standards.

What happened to developers using Metaio when it shut down?

Developers using Metaio when Apple shut it down in 2015 had no official migration path. Existing SDK installations continued to work on devices where they were already installed, but the developer portal was closed, new downloads were unavailable, and no support was offered. Most developers migrated to Vuforia for image tracking use cases. The 3D object detection capabilities had no direct equivalent until ARKit 1.5 introduced image detection in 2018 and ARKit 2.0 added 3D object detection later that year.

Was Metaio the only AR SDK with 3D object detection?

Metaio had the most capable and reliable 3D object detection implementation available on mobile at that time. Vuforia had some 3D object recognition capabilities but they were more limited. Other SDKs of the era generally focused on image target tracking rather than 3D object detection. Metaio's 3D object tracking was not the absolute only implementation, but it was widely considered the strongest on mobile hardware of that generation.