Solution
Entropy
True random numbers
Physical randomness, ecologically sourced, for workloads where true unpredictability is essential.
The problem
Most random numbers are generated by software algorithms, which makes them predictable in principle.
Pseudo-random algorithms produce sequences from a seed. That is fine for everyday use. In cryptography, simulation, fair draws or sampling, a seed or algorithm that can be inferred or reproduced weakens everything built on it.
What it does
Randomness you can measure, source responsibly and evidence.
Physical source
Measured from a natural phenomenon by high-precision sensors. Non-algorithmic: there is no seed to recreate and no sequence to reverse-engineer.
Statistically tested
Every batch is tested on a separate test server against recognised statistical standards, including the NIST test suite, before it is released.
Clear provenance
Each stream is documented: how it was sourced, processed, tested and delivered.
Ecologically sourced
Generated on a farm of dedicated, low-power mini servers powered by renewable energy.
Delivered as a service
Tested batches are held in secure cloud storage and delivered as downloads, through an API from an approved pool, or into agentic workflows via River.
A governed component
Randomness stops being a hidden assumption and becomes a visible, measurable part of your architecture.
How it works
- Source: a natural phenomenon produces unpredictable variation.
- Capture: high-precision sensors on each node in the farm measure that variation and convert it into digital form.
- Condition: each node processes its raw signal into uniform, usable output.
- Test: every batch is validated statistically on a separate test server before release.
- Store and deliver: tested batches are held in secure cloud storage, then delivered as downloads or through an API from an approved pool, each with its provenance record.
Entropy is operated by aevris and delivered as a managed service. Generation runs on a farm of dedicated mini servers powered by renewable energy, with UPS backup for continuous operation, and statistical testing runs on its own separate server. Tested batches move to secure cloud storage, which serves every delivery and acts as a buffer so supply stays continuous: you are served from storage, never from the generators themselves. Source and method details are available under NDA.
Where it delivers value
Entropy fits wherever a system's integrity rests on unpredictability that cannot be guessed, seeded or reproduced:
- Cryptographic key and token generation
- Secure communications and authentication
- Monte Carlo and scientific simulation
- Gaming, lottery and provable fairness
- Regulated draws and random selection
- AI training and evaluation where unbiased sampling matters
Ecological value
True randomness, generated on renewable energy.
Entropy is generated on a farm of dedicated, bare-metal mini servers powered by renewable energy, with UPS backup to keep them running continuously. The hardware is small, and so is the energy it takes to produce each stream.
Where it matters to you, we can document the energy source alongside each stream's provenance, to support your own sustainability reporting.
The aevris approach
We start with your workload: how sensitive it is, what throughput it needs and how it must be integrated. From there we agree the delivery, the validation evidence you need and the integration path.
Integrity at every stage: source, capture, conditioning, testing, provenance and delivery.
Talk to us about Entropy
Tell us what your workload needs and how sensitive it is. We will agree the right delivery, validation and integration.