Learns each flow’s normal
Volumes, record shapes, timing and sources are learned per flow, including the daily and weekly rhythm.
MAJRA · DATA INTEGRATION
MAJRA مجرىMove data between systems with provenance.
MAJRA is the channel data flows through. Sources are onboarded securely, collected at the edge, transformed, routed by policy and observed end to end, with the provenance of every record kept. AI watches every flow and raises the alarm when the data stops looking normal.
WHAT MAJRA DOES
Data integration.
On an open core.
- Secure source onboarding
- Edge collection with MiNiFi and Vector
- Transformation and routing by policy
- AI anomaly detection on every flow
- Provenance for every record
- Back-pressure and guaranteed delivery
- Observability across every flow
Open-source core: Apache NiFi · MiNiFi · Vector. Delivered, handed over and supported by Alin Li, and able to carry your own product name.
AI INSIDE MAJRA
AI that knows
when data looks wrong.
A pipeline can deliver every record and still deliver the wrong data. MAJRA’s AI layer watches the flows themselves.
Flags anomalies in flight
Sudden drops or spikes, silent sources, schema drift, and late or duplicated records are flagged before they reach downstream systems.
Explains with provenance
Each alert shows the baseline it broke and links to the provenance of the records involved.
Runs beside the pipeline
The models run on your own servers next to NiFi and Vector. People decide what happens next.
PART OF ONE FAMILY
Works with the
rest of the estate.
Every product is managed from the Alin Li Platform Console and shares sign-in, audit and health.
Offered
SAQR صقر
SOC and SIEM
A sovereign SIEM and XDR security operations centre on an open core.
Wazuh · OpenSearch
Offered
ATHAR أثر
Endpoint log retention
An agent per device, a collector and an index. Every event kept, and restored on request.
Vector · OpenSearch
Offered
ZIMAM زمام
SOAR and automation
Playbooks that act on what the SIEM finds.
ShuffleSTAY CURIOUS
See MAJRA on a lab estate.
A short demo with an engineer, then a time-bounded proof of concept on your own data.