ROC SDK
Multimodal
Biometrics SDK
ROC SDK packs fourteen Vision AI capabilities into one SDK —
face, fingerprint, iris, ALPR, gun detection, and more.
ROC SDK
Multimodal
Biometrics SDK
Build effortlessly with confidence. From futurized capabilities to category-defining biometrics, ROC SDK is your unified Vision AI toolkit to power it all.
Everything you need
to dive straight into dev.
INCLUDES
NIST-Ranked
ROC Biometrics
Support 500M identities
with five lines of code.
A robust, scalable biometrics foundation, ROC SDK also
powers our next-gen national security, public safety,
and ABIS platforms.
SUPPORT + SPECIFICATIONS
Made by engineers
for engineers.
Integrate our AI models into your existing workflows with speed and precision.
Spend more time building products and less time building integrations.
Personalized
support
When you win, we win. Get expert support from senior devs who actually write code.
Flexible pricing
We’re easy to work with. Choose per device, per stream, or per transaction pricing. Server or edge licensing with free dev SDK access.
Clear
documentation
Extensive docs and sample code for seamless deployment across cloud, on-premise, or edge.
Command line
interface
A robust CLI lets you build and automate systems using simple shell scripts.
Fully native solution
No internet connection required—run everything securely on your own black box.
Web API 
Deploy FR technology on your own web service with full hardware control.
GPU support
Leverage GPUs for hyper processing and maximum templatization throughput.
Video support
Process videos in real-time with a single core on mobile or server.
Lightweight
38MB base library with small template sizes for large-scale deployments of 500M+ identities.
Built on commonÂ
sense andÂ
common ground.
ROC SDK is made in America by national security experts,Â
scientists, and engineers. That’s why we’re trusted by theÂ
U.S. military, law enforcement, and leading FinTech brands.
FREQUENTLY ASKED QUESTIONS
What biometric and AI capabilities are included in ROC SDK?
ROC SDK includes 14 capabilities in one package, organized across three toolkits:
NIST-Ranked Biometrics:
- Face Recognition, Fingerprint Recognition, Iris Recognition, Tattoo Recognition
Identity Toolkit:
- Identity Verification, Age Estimation, Liveness / PAD, Face Analytics, ICAO compliance checks
Object Toolkit:
- Vehicle Recognition, ALPR, Gun Detection, Person Detection, Object Detection
All 14 share the same API interface – adding a new modality is a few lines of code, no architectural changes required. One SDK evaluation covers your entire capability roadmap.
More: roc.ai/face-recognition — roc.ai/liveness — roc.ai/object-detection
Which programming languages and platforms does ROC SDK support?
ROC SDK runs on every major development environment with no platform-specific workarounds:
Languages: C++, Java, Python, C#, Go, Node.js, Rust
Operating systems: Windows, macOS, Linux (x86-64 and ARM including Apple Silicon), iOS, Android
Deployment targets:
- Cloud infrastructure (AWS, Azure, GCP)
- On-premise servers
- Edge and IoT devices
- Air-gapped classified environments - no internet connection required
All platforms use the same API - write integration code once, deploy anywhere. A free 30-day developer license is available with same-day access to documentation and sample code.
More: roc.ai/sdk
How does ROC SDK pricing work?
ROC SDK offers three flexible licensing models designed to match actual deployment economics:
- Per-device - fixed fee per deployed device; ideal for edge kiosks, patrol hardware, and fixed terminals
- Per-stream - priced per concurrent video stream; ideal for surveillance and video analytics workloads
- Per-transaction - priced per biometric comparison; ideal for FinTech identity verification at scale
All models include a free developer SDK for evaluation and prototyping - no minimum commitment required to start building. Server and edge licensing are both available.
Contact sales@roc.ai for a custom quote based on your deployment volume and selected modalities.
Can ROC SDK run offline in air-gapped or classified environments?
Yes. ROC SDK is a fully native, on-device solution - no internet connection required and no biometric data transmitted to external servers. All processing runs entirely on local hardware.
This makes ROC SDK the preferred choice for:
- Classified government deployments where network-connected software is prohibited
- Defense applications requiring ITAR and EAR compliance
- Data sovereignty requirements where biometric data must remain within a defined perimeter
- Operational environments with unreliable or no network connectivity
Air-gapped deployment is available for all SDK modalities with identical performance to cloud-connected configurations.
More: roc.ai/national-security
How does ROC SDK scale to support hundreds of millions of identities?
ROC SDK is engineered for national-scale identity programs:
- 38MB base library - compact enough for edge devices, scalable to cloud clusters
- Fingerprint templates 160x smaller than nearest competitors - dramatically reducing storage costs at scale
- ROC's largest live production deployment: a U.S. federal government program with a 500M-template gallery at mission-critical throughput
- Horizontal cloud scaling natively supported - no reengineering as volume grows
- Sub-second search latency maintained across 100M+ gallery sizes per NIST benchmarks
From a startup's 10,000 users to a national identification program - the same SDK handles both without architectural changes.
More: roc.ai/abis
How does ROC SDK compare to building biometric algorithms in-house?
Building biometric algorithms in-house requires:
- 3-5 years of dedicated R&D before reaching competitive accuracy
- A massive, precisely annotated training dataset covering diverse global populations
- A specialized ML engineering team for continuous model improvement
- Ongoing NIST benchmarking to validate performance against competitors
ROC SDK delivers #1-ranked algorithms with days of integration time. ROC has invested 10+ years to reach top NIST rankings across face, fingerprint, iris, and tattoo - available to your team via SDK at a fraction of the cost of replicating that internally.
What are the ROC SDK base library size and performance specifications?
ROC SDK is designed to deliver maximum capability in minimal footprint:
- 38MB base library - runs on constrained edge hardware without performance trade-offs
- Fingerprint templates 160x smaller than nearest competitor
- GPU support for maximum templatization throughput in high-volume server deployments
- Real-time video processing on a single CPU core on mobile hardware
- CLI for automated batch workflows and shell script integration
- Web API option for full hardware-controlled server deployments
All specifications are verified under production conditions - not controlled lab benchmarks.
More: roc.ai/sdk — roc.ai/testing-and-standards
What documentation and developer support does ROC SDK provide?
ROC provides the documentation and support that enterprise development teams actually need:
- Language-specific documentation and sample code for all 7 supported languages
- CLI for automated workflows and batch processing
- REST API reference with interactive examples
- Direct access to senior developers - the engineers who wrote the SDK, not a first-line support queue
- A free 30-day evaluation license with same-day access to all documentation and support
Most developers have a working integration on their first day. When you ship successfully, ROC wins too - that alignment is built into how the support team operates.
More: roc.ai/sdk
How do I get started with a free ROC SDK trial?
Request a free 30-day evaluation license at roc.ai/sdk. The trial includes:
- All 14 biometric and AI capabilities - no modality restrictions during evaluation
- Complete documentation for all supported languages and platforms
- Sample code for common integration patterns
- CLI access for batch processing and automation
- Senior developer support from day one
No credit card required. No procurement cycle. Most developers have a working integration within 24 hours of receiving their evaluation license.
More: roc.ai/sdk — Contact: sales@roc.ai