An AI platform
with its own laboratory.
SynapseX is built by SoftQuantus: multi-model AI chat, a metered API platform, and the QuantumOS engine — one stack, operated under rules that fail closed.
House rules,
not house style.
Six operating principles wired into the platform itself. They are restated as commitments in our Responsible Use policy.
If a result can't carry its evidence, we don't call it a result
Every run leaves a record designed to be checked by someone who wasn't in the room. No record, no claim.
Cheapest capable probe first
Free simulators and small runs answer most questions. Expensive hardware is the last resort, never the default.
Estimate before any spend
You see the projected cost before a credit moves. No estimate, no run.
Human approval above your budget envelope
Agents work inside the envelope you set. Anything above it stops and asks you.
Guards fail closed
When a limit check cannot complete, the run is refused — not waved through.
No silent over-charges
Credits reserve before a run and settle to actual usage. Unused reserves auto-release back to you.
Who builds this.
Roytman Piccoli
Founder
Founded SoftQuantus and leads the build: the chat, the API platform, the QuantumOS engine, and the model family trained on EuroHPC's LUMI supercomputer.
SynapseX is a brand of SoftQuantus. One company builds the whole stack — the models, the metering, the lab — so every layer answers to the same rules.
Write to us.
Sales, partnerships, privacy requests, and everything else start at the same address. A person reads it.
See the stack
for yourself.
The chat and the free simulators are live today. Start there — the rules above apply from the first run.
SoftQuantus · sales@softquantus.com