Fees agreed up front. No meter once it runs.
Nxgyle bills the build; your stack carries the run. Fees follow the processes and integrations in scope. Rule steps run without a model, AI steps bill to the provider you configure, and there is no runtime license to renew.
Three engagement shapes, each priced by scope.
The core fee covers designing, testing and exporting your workers. It is sized by processes and integrations in scope, never by run counts. Managed Delivery is an add-on, and workers change only on your request. If payments to us stop, the workers keep running, because no license check exists to stop them.
Build and export
We design each worker, dry-run it and export it as a complete code project with its web service, workflow runtime, tests, container definition and README. The fee follows the processes and integrations in scope, is settled before work starts and does not move with run volume.
Managed Delivery as an option
An add-on, not a runtime dependency. We can deliver upgrades and changes for you, while your platform team runs production. Every export includes its test suite, and generated runbooks and architecture docs can be switched on at export.
Regenerate on request
Changes ship when you request them, not on a renewal cycle. With the optional upgrade path enabled, regeneration carries forward your engineers' edits and reports conflicts for review.
No rate card, by design.
Scope sets the fee: processes, integrations and deployment targets vary too much for a generic rate to be accurate. Every figure on this site follows one rule: publish only what we can defend. Your number comes from a sizing session, not a price list.
Full business applications, engineered end to end and delivered for rebranding.
Each automates one defined process under set rules and ships as source code.
Led by finance, manufacturing and healthcare, with construction, energy and telecom.
Figures reflect the delivery catalog published on this site. A digital worker is software that carries one business process from intake to outcome, following your rules. You receive full source code, not a repackaged bot or script.
No meter, license check or callback.
All three are visible in the delivered source. Your engineers can grep, build and run it on machines you control, without asking us.
No usage metering
Nothing in the code counts runs or decisions for billing, so load never changes your Nxgyle fees. Your model provider bills its calls directly, and each run's trace records its own token and cost totals.
Verify it against the export packageNo license check
The exported source has no license validation, activation call or expiry logic. Build it from the shipped container definition on hosts you choose, and nothing stops it when the contract ends.
Verify it against the export packageNo phone-home
Exported code contains no Nxgyle endpoint and no telemetry SDK. Outbound calls go only to endpoints you configure, and optional export to your monitoring stack is disabled by default.
Verify it against the export packageEach export carries a signed AI bill of materials: a signed inventory that the bundled verifier checks offline, on your machines. To inspect one before you engage, request a redacted sample through the Trust Center. Run the check yourself.
Where runtime spend comes from.
Fixed rule checks run inside the worker, so they carry no model spend. That is structural, not a discount: no model is called. Spend comes from AI steps and the fan-outs and loops around them. Each proposed design carries an estimated monthly run cost from the design-time cost model, and payback is projected once, at the planning stage.
AI steps call whichever of the eleven supported model providers you configure, local models included, with per-step model choice and fallback chains. An optional step cap and run timeout bound each run; there is no dollar cap. Each run reports its token and cost totals, and self-hosted models record no model spend. Where one decision type recurs all day, our Cognitive Engine can settle qualified routine cases with no model invocation. The engine ships separately; callers reach it over the Cognitive Studio's web API, outside Nx·Studio workers.
You get the mechanics, not a savings headline. Model your own volumes before you commit.
Our fees pay for the build, not for runtime. Usage-priced platforms gain when you depend on them; our incentives point toward a platform team that runs the code without us. Support after delivery is optional.
- Design and export Billed by Nxgyle per engagement, sized by processes and integrations, settled before work begins.scoped
- Rule steps Fixed rule checks run inside the worker without any model, adding no AI cost. Rules that cannot be expressed this way are flagged.zero
- AI steps Billed by the provider you configure, at your negotiated rates; Nxgyle adds no margin. Fan-outs and loops multiply calls, so each design shows an estimated run cost first.provider
- Hosting and runtime Your cloud or servers, built from the shipped container definition or started locally with one script. No runtime license and no callback to us.your ops
- Managed Delivery Billed by Nxgyle only when your team opts in.optional
No amounts appear here: each engagement is sized to its own scope, and a sample figure would mislead.
Does anything in the export call home or expire?
No. The export is plain source code with no Nxgyle endpoint, license check, activation call or usage meter, and it keeps running whatever happens to the contract. Outbound traffic goes only where you configure it. The signed AI bill of materials and its offline verifier let your engineers check every file.
Inspect the export filesWhat does the Nxgyle fee cover?
Designing, testing and exporting the workers under a scoped engagement. There is no runtime license and no per-run fee. Rule steps never call a model, and AI steps bill to your own provider accounts. Each design shows an estimated run cost before anything is built.
How run cost is modeledWhy are there no published rates?
Processes, integrations and deployment targets differ in every engagement, so a published rate would misstate yours. Pricing follows the same rule as every number on this site: if we cannot defend it, we leave it out. A scoping session with your architects produces the figure.
Book a scoping sessionDoes run volume change what we pay Nxgyle?
No. Run volume never changes our fees. Model usage is a separate line that your provider bills and that rises with volume. Every design estimates its monthly run cost up front, test runs log actual spend per trace event, and each production run reports its own token and cost totals.
Can our platform team run it without Managed Delivery?
Yes. Nothing in the code depends on Managed Delivery; it stays optional. Every export ships its tests, container definition and README, and generated runbooks and architecture docs can be enabled at export. You decide who maintains the workers, us included.
What exactly does engineering receive?
The complete code project (web service, workflow runtime, agents, tools, tests, container definition, README), the workflow definition, its threat model, plus the signed AI bill of materials. Exports with approval gates also bundle an approvals inbox app that engineering connects to the worker. Container-orchestration and other deployment templates are optional.
Open the Trust Center