Today we're launching TrueEntropy - a quantum random number service built from the ground up by Interlink Digital Group in Blackpool, UK. Powered by our proprietary quantum programming language QuBitLang and IBM Quantum hardware, TrueEntropy delivers genuine quantum entropy through the most elegant API we could imagine.
Why We Built This
Every random number your computer generates is a lie. Operating systems, programming languages, and cryptographic libraries all rely on pseudo-random number generators (PRNGs) - deterministic algorithms that produce sequences which look random but are entirely predictable if you know the internal state.
For most applications, that's fine. For gambling, cryptography, clinical trials, and blockchain provenance, it's not. These industries need numbers that are fundamentally unpredictable - numbers whose randomness is guaranteed by the laws of physics, not the complexity of an algorithm.
That's what TrueEntropy provides. Every random number we generate originates from a quantum measurement - a physical process that is provably unpredictable according to quantum mechanics.
How It Works
The pipeline from quantum physics to your API response has five stages:
1. QuBitLang Circuit
We define quantum random number generator circuits in QuBitLang, our proprietary quantum programming language. A Hadamard gate is applied to each qubit, placing it in an equal superposition of |0⟩ and |1⟩. When measured, each qubit collapses to a truly random bit.
2. IBM Quantum Execution
QuBitLang circuits compile down to optimised Qiskit circuits, which execute on IBM Quantum Heron processors - 156-qubit superconducting systems accessed via Qiskit Runtime. Our default backend is ibm_fez.
3. Statistical Verification
Every batch of quantum entropy passes our NIST SP 800-22 test suite before entering the pool. We run 7 core statistical tests - Frequency, Block Frequency, Runs, Longest Run of Ones, Serial, Approximate Entropy, and Cumulative Sums - at a significance level of α = 0.01. Only batches that pass all 7 tests are accepted.
4. Entropy Pool
Verified quantum random bytes are stored in a binary entropy pool file. A Python daemon monitors the pool and automatically refills it when levels drop. This pre-generation model means your API requests never wait for quantum hardware.
5. Your API Call
PHP reads from the entropy pool using atomic file locking to prevent race conditions between concurrent requests. If the pool is ever depleted, we fall back to PHP's CSPRNG with a warning header - but the daemon ensures this rarely happens in practice.
The API
TrueEntropy exposes a clean REST API with endpoints for every common use case:
- Integers - Random integers in any range, with rejection sampling for uniform distribution
- Floats - IEEE 754 double-precision random floats with 53-bit mantissa precision
- Bytes - Raw quantum random bytes in hex or base64 encoding
- UUIDs - Quantum-generated v4 UUIDs
- Bitstrings - Raw binary strings direct from quantum measurement
- Shuffle - Fisher-Yates shuffle powered by quantum entropy
- Batch - Multiple operations in a single request
Every response includes metadata: the QuBitLang circuit used, the quantum backend, and NIST verification status. Authentication uses Bearer tokens with the prefix te_live_ for production and te_test_ for sandbox.
A Quick Example
-H "Authorization: Bearer te_live_..."
That's it. Five quantum random integers between 1 and 100, sourced from IBM Quantum hardware, verified by NIST statistical tests, delivered from a pre-generated entropy pool.
What's Next
TrueEntropy is live and accepting registrations. The Free tier gives you 1,000 requests per month - enough to explore every endpoint and integrate quantum entropy into your projects.
In the coming weeks, we'll publish deep dives into QuBitLang's compiler architecture, our NIST testing pipeline, the entropy pool system, and practical tutorials for regulated industries. If you're building something that demands true randomness, we'd love to hear from you.