Sell inspection without building
the inspection software.

xis.ai is a no-code AI vision platform that runs on standard industrial hardware. If you already sell to manufacturers — as an integrator, a hardware manufacturer or an advisor — it is designed to sit inside what you deliver rather than compete with it.

Request partnership

Partner tracks

Four ways
to work with us.

These are different relationships with different economics, so they are set out separately rather than bundled into one programme.

  • Delivery

    System integrators, machine builders & automation companies

    You already build, commission or automate production lines. xis.ai supplies the vision layer so you are not writing or maintaining detection software, and your engineers configure it themselves rather than subcontracting every change back to a vendor. Where you run your own automation platform or app ecosystem, xis.ai can sit inside it.

    You bring

    Customer relationships, mechanical and electrical integration, commissioning and local support.

    We bring

    The platform, training for your engineers, and our team alongside you for the first models on a new account.

  • Compute

    Industrial PC & edge hardware manufacturers

    xis.rt runs on an industrial PC with an NVIDIA GPU at every inspection point. Each deployed line is a unit of GPU-class edge hardware, and a validated configuration makes your platform the straightforward answer when a customer specifies AI inspection.

    You bring

    Edge compute platforms, your channel, and customers already asking about industrial AI workloads.

    We bring

    Validation of your configurations, a reference architecture, and joint technical material your sales team can use.

  • Imaging

    Camera, optics & lighting manufacturers

    We use standard industrial cameras and interfaces rather than proprietary hardware, so your products work without driver development on either side. For applications where rule-based tools stall — reflective surfaces, variable product — the deep-learning layer is what makes your imaging sellable.

    You bring

    Cameras, optics, lighting and the application expertise that decides whether an inspection is feasible at all.

    We bring

    The inference layer, validated support for your interfaces, and joint application notes for the cases rules cannot solve.

  • Advisory

    Consultants & engineering advisors

    You are in the room when a manufacturer decides whether to automate inspection. Our feasibility check gives you something concrete to put in front of a client — a model trained on their own product, with an honest answer, including when the answer is that it will not work.

    You bring

    Trusted access at the point where inspection strategy is set, and knowledge of the client's process.

    We bring

    The feasibility assessment on the client's own images, technical input during scoping, and support through the pilot.

What you would be working with

The technical surface,
in full.

Most partner conversations stall on unanswered technical questions. These are the answers, stated once.

Shared by both

xis.dev

Model development

Models

Quality

Line 1

xis.rt

Line 2

xis.rt

Line 3

xis.rt

Results

xis.dashboard

Every line, plant and country

Safety

Production floor

existing cameras

Warehouse

existing cameras

Loading bay

existing cameras

Streams

xis.safety

Runtime and dashboard in one

Hardware per inspection point

  • CameraStandard industrial cameras and interfaces
  • LightingControlled illumination, specified per application
  • ComputeIndustrial PC with an NVIDIA GPU
  • Existing equipmentRetrofits onto cameras, lighting and IPCs already on the line in most cases
  • Proprietary hardwareNone required

Integration & deployment

  • PLC and MESOPC-UA, MQTT and PROFINET
  • Line signallingReject, alarm or line halt in real time
  • Inference throughput150 fps; achievable line rate confirmed per application
  • ConcurrencyMultiple products and multiple defect types on one runtime
  • Data residencyEntirely on customer infrastructure; no vendor cloud
  • ScalingOne runtime per inspection point; dashboard added when a multi-site view is needed

What every partner gets

Enablement, not a logo
and a login.

The same handover our direct customers receive, plus what a partner needs in order to sell and deliver without us in the room.

  • 01

    Platform training

    Working sessions with our engineers so your team can build, validate and deploy models independently.

  • 02

    Documentation

    A full platform manual and training videos, so new people on your side can be brought up without scheduling us.

  • 03

    Technical support

    Direct access to our engineers during scoping, feasibility and deployment, with remote support after go-live.

  • 04

    Model building as a service

    Where you would rather not build models yourself, we collect the data and build them — permanently or just for the first deployments.

  • 05

    Feasibility assessments

    We run the assessment on your client's images and give you an honest answer to take back, including when it is no.

  • 06

    Joint material

    A reference architecture and application notes, built with you around your first joint deployment.

Before any agreement

Bring us a
real application.

Yours or a customer's. We train a model on those images and come back with an honest read on whether it works — which is a faster way to judge a partnership than a presentation.

Run a feasibility check

Partner questions

What partners ask before the first call.

  • No. xis.dev is a no-code environment designed for engineers rather than data scientists, and we train your team on it. Where you would rather not build models at all, we can collect the data and build them for you — either permanently or only for the first deployments.

Still deciding

The one question this page can't answer:
does it work on your defect?

  • Start here

    Send us your images

    Share examples of the defect you care about. We train a model on your product and come back with a straight answer — including if it's no.

    Start a feasibility check
  • Or ask directly

    Talk to an engineer

    Thirty minutes with someone who builds this. Bring your line speed, your defects and your integration constraints.