What makes Lokey a true random number generator (TRNG)?
Lokey uses a dedicated hardware secure element to produce true entropy.
Lokey
Hardware-grade True Randomness as a Service
Most “random” numbers on your systems come from algorithms that can, in principle, be predicted. Lokey streams true hardware randomness from physical noise through an open-source TRaaS API, so your cryptographic keys, AI models, and critical systems start from real unpredictability.
Open source, EU-hosted, no cookies or trackers — verifiable randomness for mission-critical systems.
Proof You Can Audit Today
We publish test reports, run without trackers, and keep every line of code open so your team can verify our randomness.
45 / 46
Dieharder statistical tests passed
View full results →
0
Cookies or trackers used
Privacy-first by design
100%
Open-source codebase
No black boxes, community-audited
Why Pseudo-Random Isn’t Enough for Critical Systems?
True vs. Pseudo-Randomness in Practice
Computers default to pseudo-random number generators (PRNGs)—algorithms that look chaotic but are fully determined by a seed. If an attacker predicts or steals that seed, every subsequent “random” value becomes predictable. True randomness instead comes from messy physical processes, like electronic noise, that have no underlying pattern. For cryptography, secure communications, regulated AI workloads, or national infrastructure, that difference matters. As quantum computing advances, purely algorithmic randomness becomes even easier to exploit, while physics-based entropy stays resistant.
- Encryption & Key Management: Physical entropy keeps key material unpredictable even under quantum-capable adversaries.
- Machine Learning: Truly random weight initialization helps models avoid symmetrical traps and local minima.
- Cloud & Multi-Tenant Security: Hardware randomness enforces isolation and secrets across shared infrastructure.
Our Mission
Lokey is an open-source foundation for True Randomness as a Service (TRaaS). We democratize access to hardware-generated entropy with transparent code, published test data, and commercial support only where it adds value.
Commercial Strategy
We deliver value-added services around open hardware and software: managed entropy devices, TRaaS endpoints, compliance support, and custom integrations for high-stakes customers.
How Lokey Works (Hardware-Backed TRaaS)
1. Physical Entropy Capture
Microchip’s ATECC608B amplifies microscopic thermal noise—tiny electrical fluctuations that are fundamentally unpredictable.
2. Open-Source Supervision
Our lightweight firmware and health tests continuously monitor the entropy stream, catching bias or hardware faults in real time.
3. Fortuna Cryptographic Mixing
We feed the raw bits into Fortuna, the robust RNG architecture by Bruce Schneier and Niels Ferguson, to harden output against attacks.
4. TRaaS Delivery & Auditing
Entropy streams are exposed via API endpoints, CLI tooling, and on-device services—alongside published Dieharder scores and visual analyses for verification.
Where Lokey TRaaS Fits in Your Stack
Four sectors already rely on true randomness to remove single points of failure in cryptography, compliance, and AI. Expand each card to see the workflows Lokey secures.
Governmental Sector
Integrity and verifiability are non-negotiable for digital services such as identity, voting, notarization, and public lotteries. Lokey provides auditable entropy sources so every key, token, or draw can be challenged and verified.
- Digital signature keypairs and certificate issuance
- Secure digital voting or lottery randomness disclosures
Banking & Finance
Financial infrastructure needs defensible randomness for transaction IDs, one-time passwords, customer onboarding, and secure device provisioning. Lokey adds hardware entropy without ripping out existing HSMs.
- Fortify HSM or KMS seed pools with verifiable entropy
- Tamper-evident randomness logs for audits and regulators
Cloud Providers
With multi-tenant environments under regulatory pressure, cloud platforms need hardware randomness to isolate customers, encrypt workloads, and issue API credentials that cannot be replayed.
- Per-tenant entropy feeds for VM, container, and serverless workloads
- Secrets management services seeded by auditable randomness
AI & Machine Learning
From reproducible research to safety-critical autonomous systems, AI teams depend on good randomness for weight initialization, data shuffling, and privacy-preserving techniques.
- Truly random seeds for training and differential privacy mechanisms
- Entropy for simulation, Monte Carlo sampling, and reinforcement learning
Randomness Validation & Testing
Dieharder & Statistical Batteries
Lokey’s output passes 45 out of 46 Dieharder tests—the industry standard battery derived from George Marsaglia’s original suite and the NIST STS. A 1–5% failure rate is expected for a healthy TRNG; we publish full logs for community review.
Read Dieharder reportVisual Analyses
We render 512×512 Data Bitmaps, frequency histograms, and autocorrelation plots. Healthy randomness looks like TV static (no patterns), flat histograms (no bias), and near-zero autocorrelation (no repeated structures).
- Mean intensity ≈ 127.5 and standard deviation ≈ 73.9 confirm uniform coverage of byte values.
- Pattern score and max autocorrelation stay low, warning us if a hardware component drifts.
Frequently Asked Questions
Frequently Asked Questions
This FAQ covers what we mean by randomness and entropy, how Lokey generates and validates its numbers, and how you can deploy the service inside secure or regulated systems.
Why do I need hardware entropy instead of relying on my OS?
Hardware entropy provides physical unpredictability unavailable to software PRNGs.
How does Lokey ensure quality and safety of the random data?
Lokey uses both hardware tests and cryptographic strengthening.
Can I self-host Lokey inside secure or regulated environments?
Lokey is designed for self-hosting in strict environments.
What tests does Lokey use to validate randomness?
Lokey uses industry-standard test suites and visual diagnostics.
What is the Data Bitmap and why is it useful?
Bitmap visualization reveals bias quickly.
Can occasional Dieharder failures indicate a broken TRNG?
Occasional failures are normal; consistent patterns are not.
Is Lokey open source?
Lokey is fully open source.
How Lokey Evolves from MVP to Global Service
Phase 1: MVP & Device Genesis
- Finalize the MVP of the TRaaS API
- Develop and complete an open-source hardware device and ecosystem under €50
- Prepare for external security audits and compliance checks
Phase 2: Product Validation
- Engage with early adopters in governmental, banking, and cloud sectors
- Refine the TRaaS solution based on real-world feedback
- Launch a dedicated developer portal and documentation hub
Phase 3: Scaling for Commercial Use
- Deploy geographically distributed entropy sources for global reach
- Establish a formal governance model with community and partner input
- Develop and launch commercial-grade support and specialized services
Built and Hosted in the EU for GDPR-First Security
Lokey is developed and operated entirely within the European Union. Your data remains under EU jurisdiction, protected by GDPR and shielded from extraterritorial demands such as the U.S. Patriot Act or CLOUD Act. If regulatory sovereignty matters to you, Lokey keeps your randomness pipeline compliant by default.
The Team Behind Lokey
Alex Popatanasov
Research & Cryptography
Researcher focused on cryptography, algorithms, and Linux—driving Lokey's technical innovation and auditability.
Stanisa Obradovic
DevOps & Integrations
Runs infrastructure, automation, and third-party integrations—scaling Lokey reliably for enterprise workloads.
Partner with Lokey to Shape the Randomness Layer of the Internet
We are building an ecosystem around true randomness — and we can’t do it alone. We’re looking for visionary organizations, auditors, and technology integrators to help us define the next-generation randomness infrastructure.