We price the build. Never per decision.
Your AI workforce should be an asset, not a subscription. Fees are agreed against a defined scope before work begins. There is no per-decision meter or runtime license; your model provider bills AI usage directly.
Spend tied to scope, not to volume.
Finance leaders ask what will grow as volume grows. Our fees will not. They cover the design and delivery of your workforce, priced on the processes in scope and the systems we integrate. Managed Delivery is optional, and upgrades happen only when you commission them. If you stop paying us, nothing stops working, because the code is yours.
Design & delivery engagement
A defined project with an agreed fee: we design, validate and deliver your digital workers. Its price follows the processes in scope and the systems we integrate; decision volume never moves it.
Managed Delivery — your choice
Ongoing support you choose, not a dependency we build in. Retain us for upgrades and new changes as your own teams operate the workers, or hand everything over. Tests ship in every export; runbooks and dashboards for monitoring can be added.
Upgrades on your schedule
Investment follows business change, not a subscription cycle: commission a regenerated worker when a process changes. An optional upgrade tool, off by default, merges your engineers' edits into the new version and flags conflicts for review.
There is no price list here, and that is a deliberate choice.
Every engagement is scoped to your processes and systems, so a list price would mislead more than it informs. The claims policy behind every number on this site applies here too. We set your figure in a scoping session rather than publish a number we cannot stand behind.
Complete business applications, delivered to run under your own brand.
Owned software assets, each running one defined process by your rules.
Across financial services, healthcare, energy, manufacturing, logistics, legal and automotive.
Totals are derived from the delivery records this site publishes. Every digital worker is owned outright by your enterprise: software for one defined business process, run within the rules you set, with the source code in your hands.
Three vendor ties you never carry.
Each absence is a property of the code you receive, not a contract term that can lapse. Your security team can verify all three on your own hardware.
No per-decision fees
Metered vendors turn growth into a rising bill. Nxgyle has no meter, so volume never raises what you pay us. AI calls are billed separately by your model provider, and that cost does grow with volume.
Verify it: what every export containsNo runtime licensing
Each export is complete source code on your infrastructure. No license key, activation check or subscription keeps it running, so continuity never depends on renewing a contract with us.
Verify it: what every export containsNo telemetry to Nxgyle
A delivered worker sends nothing to Nxgyle: no usage data to report and no switch for us to turn off. A signed bill of materials lists every component for your security team to verify.
Verify it: what every export containsEvery export includes a signed bill of materials that your security team can verify offline, without contacting us. A redacted sample is available on request through the Trust Center. The conclusion is yours to draw.
The economics of ownership.
Wherever possible, the business rules you state become deterministic checks, and those carry no AI cost by design. This is not a discount: no AI model is called. Every design also shows an AI-estimated running cost and hours saved, and the generation plan adds a payback estimate. Finance can compare the options before any budget is committed.
The AI judgment that remains runs on the model accounts you choose, including fully local models. Each run can also be capped by steps and duration. You pay your model provider directly, at your own rates, with no Nxgyle meter on top. For agent systems that make the same decisions all day, our Cognitive Engine removes proven routine decisions from the AI bill. It runs as its own product, accessed through its Studio's decision API rather than built into Nx·Studio workers.
We publish the mechanism, not a marketing percentage, so your finance team can model its own costs before signing anything.
We are paid to design your workforce, not to meter it. A metered vendor earns more as your dependence grows; our model points toward your independence. Ongoing support stays optional, on your terms.
- Design and delivery Paid to Nxgyle. Priced on the processes in scope and the systems we integrate, and agreed before work begins.agreed
- Rule-based steps No AI cost where rules run as deterministic checks, which never call a model. Rules that cannot be encoded are flagged.zero
- AI judgment steps Billed by your model provider at your rates, with no Nxgyle meter on top. Each design shows an AI-estimated running cost before you build.direct
- Running the workers Your cloud or data center, at your cost. No runtime license and no calls home.in-house
- Ongoing support Managed Delivery, paid to Nxgyle only if you choose it.elective
No figures appear here, by design. Engagements are scoped, not metered, and a generic number would not apply to your scope.
What happens if we stop paying Nxgyle?
Operations continue, because the code is yours. Your workers are standard source code on your infrastructure. They never contact us, and there is no meter to switch off. Every export includes a signed bill of materials your security team can verify offline.
See what every export containsHow is pricing structured without a meter?
You pay for design and delivery, not for usage. There is no per-decision fee and no runtime license. Wherever possible, your business rules run as deterministic checks that never call a model. The remaining AI judgment runs on your own model accounts, and each design shows its AI-estimated running cost before you build.
The platform's cost mechanismWhy are no prices published?
Each engagement is scoped to your processes and the systems we integrate, so a generic figure would not reflect your situation. Our claims policy applies here too: if we cannot publish a number honestly, we do not publish it. A scoping session produces a figure for your own scope.
Request a scoping sessionDo our fees to Nxgyle rise as volume grows?
No. There is no Nxgyle meter, so growth never raises what you pay us. The AI judgment your workers still need is billed by your model provider, and that cost does grow with volume. That is why each design shows an AI-estimated running cost before you build, and test runs measure what each run actually cost.
Is Managed Delivery mandatory?
No. Managed Delivery is optional, a service you choose rather than a dependency we create. You can retain Nxgyle for upgrades or run everything with your own engineers. Either way, every export includes tests, and runbooks and monitoring can be included. Ownership means you decide who maintains it, including us, on your terms.
What do we own at the end of an engagement?
All of it: the source code, its blueprints and tests, an inbox app for reviewers that your engineers connect, and optional monitoring dashboards and deployment files. Everything runs on your infrastructure, with the cloud and models you choose, no license keys and no runtime dependency on Nxgyle.
Visit the Trust Center