Pricing
One subscription. Sized to your estate. Anchored on outcomes.
We do not publish a price list, and this page tells you why, honestly. What we can tell you before a single conversation: how the model works, what moves the number, what a POC costs you commercially, and what we need from you to quote accurately.
- One subscriptionThe whole platform: Sentinel, ProcBot, Sherlock, ten pillars, Agent Studio. No per-module surprises.How it is packaged
- Zero per-seat feesSeats never limit who can use the platform; the value is work delivered, not logins held.What we refuse to meter
- Estate-sizedPriced on what the platform manages: devices, nodes, services, telemetry, deployment model.What the number scales with
- Outcome-anchoredThe conversation is about the work taken over, such as the share of L1 volume automated.How we negotiate
The model
How Opstral pricing works
Three properties define the commercial model, and each one is a deliberate choice rather than an accident of packaging.
One platform subscription
A single subscription covers the platform: Sentinel AI, ProcBot, Sherlock, the ten Ops pillars, Agent Studio and the connector library. There is no per-seat licensing, because a platform whose value is work delivered should never make you count who is allowed to look at it, and no per-module surprise pricing waiting after signature.
Sized to your estate
The number scales with what the platform actually manages: devices, nodes and services under management, telemetry volume, the integrations in scope, and the deployment model. A 200-node estate and a 27,000-device carrier network are not the same product experience, and they are not the same price.
Anchored on outcomes, not seats
Opstral is governed Service as Software: the platform delivers completed work, resolved incidents, executed procedures, verified outcomes. The commercial conversation is anchored the same way, on the scope of work the platform takes over, such as the share of L1 volume it automates, rather than on how many humans hold a login.
Cost drivers
What moves the number, in both directions
When we scope an estate, these are the variables doing the work. Knowing them ahead of the call shortens the path to a real quote.
Estate size and telemetry volume
Devices, nodes, services and pipelines under management, and the volume of signals they emit. The largest single factor.
Deployment model
SaaS, private cloud, or on-prem and air-gapped inside your perimeter. Sovereign deployments carry real delivery effort, and the price says so.
Scope of autonomy
How much work the platform takes over: monitored-only estates price differently from estates where governed automation carries the routine tier.
Managed Service Ops
When our engineers operate the estate with you, the subscription includes an operations commitment, not just software.
Integration surface
The count and complexity of systems being connected: observability stacks, ITSM, network elements, custom estates.
Rollout pace
Phased rollouts shift spend to match value delivery. Starting with one domain and expanding costs differently from switching on everything at once.
Deployment
Deployment models carry different cost profiles, and we say so
The platform is the same in every deployment model. The delivery and operating effort is not, and pretending otherwise would just move the surprise to the negotiation.
SaaS
The reference cost profile: fastest to stand up, no infrastructure on your side, continuous updates. LLM inference runs on hosted providers such as OpenAI, Gemini or Groq, behind the Data Governance Fabric.
Private cloud
Runs in your cloud tenancy. Adds deployment and environment effort on both sides; your infrastructure costs stay yours.
On-prem and air-gapped
The full platform inside your perimeter, LLM inference on your hardware included, updated by signed offline artifacts. The highest delivery effort in the family, priced accordingly, and for sovereign and regulated environments it is the only honest option.
Plan the GPU budget with us before you commit: air-gapped inference needs customer-provided GPUs able to host a 120B-class open-weight model, on throughput-class hardware such as NVIDIA H100 or H200, sized to your concurrent load. We size it with you during scoping.
Managed Service Ops
Our engineers operate the estate with your team. The subscription includes the operations commitment, which is why it is priced as a service relationship rather than a software license.
The POC
The POC is a paid engagement, and that is a feature
We recommend starting with a scoped proof of concept before any longer-term commitment, and we charge for it as a standalone engagement. That is deliberate, on both sides.
Why paid
A paid POC keeps both sides serious: you commit real evaluation attention, we commit real engineers against your real environment and constraints. Free pilots produce polite demos; scoped engagements produce evidence.
What you keep either way
The deliverable is a written assessment you keep whether or not you proceed: the evidence-linked root cause of the problem we scoped, the blast-radius map, ranked automation candidates, and an honest go or no-go, including ours.
How it is scoped
Jointly, against your top recurring problems, with read-only access first and every write action approval-gated. The POC mechanics are described in the FAQ; the commercial shape is a fixed, agreed fee for the agreed scope.
No price list
Why there is no number on this page
Enterprise estates differ by orders of magnitude, so any single published number would be wrong in one of two directions: misleadingly small for a carrier network, or padded enough to scare off a mid-size estate. We would rather scope first and quote a number we can defend line by line. What we build with every customer during scoping is a tailored ROI model grounded in your own incident data and operational costs, so the business case is your numbers, not ours.
Pricing questions, answered plainly
How is Opstral priced?
One platform subscription, sized to your estate: devices, nodes and services under management, telemetry volume, integrations and deployment model. The conversation is anchored on the work the platform takes over, not on seats.
Is there per-seat licensing?
No. Seats never limit who can use the platform. A platform that delivers completed work should not charge you for the number of people allowed to watch it work.
Does the deployment model change the price?
Yes, and we say so plainly. SaaS is the reference profile; private cloud adds environment effort; on-prem and air-gapped carry the highest delivery effort, LLM inference on your hardware included; Managed Service Ops includes an operations commitment.
Is the POC free?
No. The POC is a paid, standalone engagement with a fixed fee for an agreed scope. You keep the written assessment either way: root cause, blast-radius map, ranked automation candidates, and an honest go or no-go.
Is there a minimum contract size?
Opstral is built for enterprise environments with meaningful operational complexity. We define the commitment with each customer around use case, scope and rollout pace rather than enforcing an arbitrary floor.
What do you need from us to quote?
A rough estate picture: how many devices, nodes and services, the observability and ITSM tools in place, the deployment model you require, and the operational problems you most want off your team's plate. One working session usually covers it.
Get a number you can defend internally
One working session to map your estate and your top recurring problems, then a scoped quote and a tailored ROI model built on your own incident data.