Your own applications on the same infrastructure.
Atlas, Socius and Chronos are modules — built on an SDK that is also open to your team. Your own applications automatically inherit sovereignty, ontology, audit trail and explainability. No fork, no parallel infrastructure.
Five primitives with which you can build anything.
The same building blocks with which we built Atlas, Socius and Chronos — documented, versioned, usable by your teams.
Connect any source
Adapter pattern for REST, gRPC, S3, OPC-UA, GIS, databases. Your connectors are first-class — no special status compared with the built-in ones.
Declarative pipelines
Source → inference → ontology → output, as code. Versionable, testable, reproducible — from the notebook to regular operation.
Plug in your own models
Register your own ML models in the Cortex registry — with confidence tracking, versioning and audit trail like built-in models.
Extend the domain
Add your own concepts (actor subtypes, events, relationships) to the shared ontology — compatible with Atlas/Socius/Chronos.
Interfaces that fit Makor
Components for maps, graphs, time series, audit displays. You build interfaces for your workflows, not for ours.
Same infrastructure
Custom applications run on the same on-prem/private-cloud infrastructure as the modules. No second stack, no second audit pipeline.
// Custom connector — bind any source to Axis from makor.sdk import Connector, Flux, Cortex class RadarFeed(Connector): schema = "axis://radar.v1" def stream(self): yield from self.poll(interval="5s") // Compose pipeline → Kontinuum pipeline = (Flux("radar.lagebild") .source(RadarFeed()) .infer(Cortex.model("anomaly.v3")) .annotate(ontology="verkehr") .publish("lagebild.verkehr.tiles")) pipeline.deploy(env="on-prem", audit=True)
One connector, one pipeline, one model — and a custom application with audit trail and explainability is done. Schematic.
What teams build themselves.
Your domain has terms that no standard module knows. You model them in the ontology and build the interface your team really needs.
Your research team has a model nobody else has. You plug it in as a Cortex model — with the same audit pipeline as built-in models.
Map recurring evaluation and approval processes as Flux pipelines — versionable, testable, with an audit trail per run.
Build data products for other departments without demanding a data warehouse — results as an API, tile service or notebook output.
Your team builds. We provide the backbone.
In the pilot your engineers get SDK access, training and pair sessions with our team. You keep code, data and audit logs.