Luca Passani, CTO @ ScientiaMobile, January 14, 2026
We have something big to announce: The WURFL Device Detection module for Prebid server is now officially available on Prebid.org (of which ScientiaMobile is a proud member). The documentation of the WURFL PBS Module can be found here.
In case you are in the dark about programmatic advertising, Prebid Server (often abbreviated as PBS) is the server-side version of Prebid’s header bidding technology. While Prebid.js runs auctions in the user’s browser, Prebid Server moves the entire header bidding auction process to the cloud/server. This allows publishers to:
- Reduce page load latency (the browser can focus on rendering content instead of making dozens of bid requests).
- Scale to more bidders without browser limitations.
- Support environments where client-side JavaScript isn’t feasible (e.g., AMP pages, mobile apps, video players, or CTV/digital out-of-home).
PBS is a backend auction engine that collects bids from multiple SSPs/ad exchanges on behalf of the publisher, picks the winner(s), and returns the result — improving performance, revenue potential, and user experience compared to pure client-side header bidding. Many companies offer managed Prebid Server hosting, or, if you are a large publisher, you can self-host your own PBS instance—it’s open source.
Figure 1: High-level representation of Prebid Server
Device Intelligence and Header Bidding
The new WURFL module for Prebid Server (available for both the Java and Go forks) brings device detection directly into PBS auctions.
Unlike the WurflRTD module (which focuses on client-side Prebid.js to enrich OpenRTB device data), this PBS module runs on the server. It processes incoming User-Agent strings (and Client Hints, if available) during the auction to populate the OpenRTB request with detailed device capabilities, such as precise device model, form factor, OS details, device price and release date, and more.
Key values include:
- Better targeting and bidding precision for SSPs/DSPs (e.g., higher bids for premium devices or specific models).
- Improved ad relevance and user experience in server-side environments (PBS+Prebid.js hybrids, AMP, apps, CTV, video, etc.) where client-side JS detection is limited or unavailable
- Reduced reliance on potentially inaccurate or stripped-down User-Agents post-privacy changes.
- Revenue uplift through more sophisticated device-based floor rules, creative selection, or buyer optimization.
- Crawler and Bot detection.
These aren’t theoretical benefits. A/B testing with partner publishers has demonstrated consistent CPM lifts of 3-6% when WurflRTD enrichment is present, with sophisticated DSPs (like The Trade Desk) showing double-digit percentage improvements. Notably, iOS traffic—where accurate device identification is hardest—shows the strongest gains.
This module enhances PBS’s strength in low-latency, scalable server-side header bidding by making rich, real-time device intelligence natively available. By enriching every OpenRTB bid request with comprehensive device information, bidders can immediately leverage for smarter targeting and optimization decisions. In the era of signal loss, this is particularly valuable.
WURFL Properties
The WURFL Device Detection Module automatically populates critical device fields in the OpenRTB device object:
- make: Device manufacturer (Apple, Samsung, OnePlus, Vizio, Google, etc.)
- model: Specific device model (iPhone 15 Pro, Galaxy S24, Pixel 9)
- os / osv: Operating system and version
- devicetype: Form factor classification (mobile, tablet, desktop, CTV)
- h / w: Screen dimensions in pixels
- ppi / pxratio: Pixel density metrics
- js: JavaScript support indicator
When extended capabilities are enabled, the module also enriches device.ext.wurfl with the WURFL device ID and additional device properties—from hardware specifications to browser capabilities.
Fully On-Premise: Zero Latency, Zero Cloud Dependencies
The WURFL PBS module operates entirely on-premise. This means zero external API calls during auction processing—device detection happens locally with sub-millisecond latency. For Prebid Server operators, this translates to no added network overhead, no external service dependencies, and no auction timeout risks.
The module is available for both PBS-Java (using WURFL OnSite Java API) and PBS-Go (using WURFL Infuze), giving hosting companies full flexibility in their deployment architecture.
PBS + WurflRTD: Better Together
While the WURFL PBS module delivers top-notch device detection for server-side auctions, it truly shines when combined with our WurflRTD client-side module.
The iPhone Challenge: Solved
Anyone who has worked with device detection knows the iPhone problem: Apple’s privacy-focused design means that User-Agent strings alone cannot distinguish between iPhone models. An iPhone 15 Pro Max and an iPhone 12 Mini present virtually identical server-side signatures. WurflRTD solves this. Running in the browser, WurflRTD can execute client-side benchmarks and leverage device-specific characteristics to accurately identify the exact iPhone or iPad model—information that simply isn’t available server-side.
Client Hints: Captured Where They Live
Google’s User-Agent Client Hints (UACH) is the new way Chrome (and Chromium-based browsers) signal their device type. These hints—including the critical high-entropy values that reveal device model, platform version, and architecture—are sent by the browser and accessible client-side. WurflRTD captures and resolves Client Hints before they’re ever needed, ensuring that Android device detection achieves maximum accuracy. The resolved WURFL data is then available for server-side auctions through the enriched bid request.
The Synergy
When publishers deploy both WurflRTD and route their bid requests through a PBS instance with WURFL support, WURFL delivers the most comprehensive device intelligence possible for programmatic auctions. To reiterate, the PBS module works perfectly standalone, but with WurflRTD in place, you’re getting the complete picture.
Figure 2: High-level representation of Prebid Server augmented with WurflRTD for Prebid.js.
Who is this for?
This module is meant for companies that run a Prebid Server instance, be they large publishers, SSPs or Publisher Monetization Platforms (AKA Yield Management Platforms) who provide programmatic infrastructure as a managed service.
Getting Started
The WURFL Device Detection Module is available now for Prebid Server. Full documentation is available at:
For PBS-Java deployments, the module integrates with the WURFL OnSite Java API. For PBS-Go, it uses WURFL Infuze. Both require a WURFL license from ScientiaMobile—evaluation licenses are available for testing.
FAQ: The WURFL Module for Prebid Server
- Q: Does this module require an external API call during the auction?
- A: No. The WURFL module runs locally within your Prebid Server instance (using WURFL OnSite for Java or WURFL Infuze for Go). This ensures sub-millisecond lookups with zero network latency.
- A: No. The WURFL module runs locally within your Prebid Server instance (using WURFL OnSite for Java or WURFL Infuze for Go). This ensures sub-millisecond lookups with zero network latency.
- Q: How does this interact with User-Agent Client Hints (UACH)?
- A: The module is fully UACH-aware. It automatically parses the high-entropy client hint headers sent by modern browsers to ensure accurate detection of Android versions and hardware, even when the standard User-Agent string is frozen.
- A: The module is fully UACH-aware. It automatically parses the high-entropy client hint headers sent by modern browsers to ensure accurate detection of Android versions and hardware, even when the standard User-Agent string is frozen.
- Q: Can I use this in a multi-tenant PBS environment?
- A: Yes. The module is designed to be efficient enough for high-volume, shared clusters while maintaining the security and isolation required for multi-tenant setups.
- A: Yes. The module is designed to be efficient enough for high-volume, shared clusters while maintaining the security and isolation required for multi-tenant setups.
- Q: What happens if a device is not recognized?
- A: The module will fall back to “family default values” for specific fields, ensuring the auction continues without errors while still passing meaningful properties to downstream bidders.
A Complete Ecosystem Solution
With official modules now available across Prebid.js (WurflRTD) and Prebid Server (WURFL Device Detection Module), ScientiaMobile offers a device detection solution purpose-built for the complete Prebid ecosystem. Device intelligence is no longer a nice-to-have optimization. It’s infrastructure.
Please contact us for more information and pricing.