Log in Contact us
Decision intelligence platform

Turn fragmented data into actionable intelligence

Ingest CSV, PDF, JSON, IoT, and more. Validate and unify. Build knowledge graphs. Search, forecast, and decide — with evidence.

12.4k Events / day
847 Entities linked
23 Active alerts
Request a demo View architecture
Denerio · command center Live pipeline

Sources → data lake → graph intelligence → action

  • 01 INGEST history_2025.csv → raw/ now
  • 02 VALIDATE Badge #4821 → Person 1s
  • 03 GRAPH ENTERED Zone B · 14:02 3s
  • 04 ANALYZE Graph + vector match 6s
  • 05 ACT Case opened · Zone violation 9s
Multi-format ingestion Audit-ready provenance Hybrid AI (cloud + local path) Evidence-first decisions

How it works

From raw inputs to decisions — in four stages

Denerio follows the same integration-first model used by enterprise intelligence platforms: trust the data, model the world, then analyze and act.

  1. 01

    Connect & ingest

    APIs, databases, files, documents, IoT streams, and human reports land in a governed data lake with source registry and lineage.

  2. 02

    Validate & unify

    Schema checks, repair, deduplication, and PII controls. Bad data is quarantined — never silently dropped.

  3. 03

    Graph & search

    Entities and relationships become a temporal knowledge graph. Hybrid keyword, vector, and graph search power investigations.

  4. 04

    Detect & act

    Rules, risk scoring, and AI-assisted reasoning feed alerts, cases, and audit-ready outcomes.

Product

Everything you need to go from raw data to decisions

A unified platform for ingestion, ontology, graph intelligence, and operational workflows.

⬡

Universal ingestion

Accept CSV, Excel, PDF, images, JSON, SQL dumps, and streams. Raw evidence stored with validation and quarantine.

◈

Knowledge graph

Temporal relationships across people, assets, zones, and events. Reconstruct timelines and link analysis with provenance.

◎

Hybrid search & AI

Keyword, vector, and graph search combined. Hugging Face inference for extraction and reasoning — routed through a secure gateway.

⚡

Alerts & workflow

Rule-based detection, risk scoring, cases, and outcomes. Close the loop from insight to action with full audit trails.

Use cases

Built for security and operations teams

Start with physical security and facility intelligence. The same platform extends to supply chain, operations, and risk domains.

Investigation

Timeline reconstruction

Who was in Zone B between 2:00 and 2:30 PM when the alarm fired?

Combine badge scans, CCTV metadata, and incident reports into one temporal view with full provenance.

Link analysis

Person & vehicle linkage

Which vehicles entered Gate 3 before this person accessed a restricted area?

Graph relationships across access control, ANPR, and IoT sensors — not disconnected spreadsheets.

Operations

Alert triage & cases

Is this tailgating, loitering, or a false positive?

Rule-based detection with evidence attached. Open cases, assign analysts, and close the loop with audit trails.

Intelligence

Multi-source fusion

What patterns emerge across 12 months of history and live feeds?

Ingest archived CSV dumps, news, and live streams into one pipeline — raw evidence preserved, insights layered on top.

Ontology

A shared language for your data

Every source maps to canonical objects and relationships — the foundation for graph search, rules, and AI.

Core objects

  • ◉ Person
  • ◉ Vehicle
  • ◉ Zone
  • ◉ Event
  • ◉ Sensor
  • ◉ Alert & Case

Temporal relationships

  • ENTERED Person → Zone
  • EXITED Person → Zone
  • PASSED_THROUGH Vehicle → Gate
  • TRIGGERED Sensor → Event
  • ASSOCIATED_WITH Person → Vehicle
  • REPORTED_BY Event → Person

What connects in

IoT People counters, motion, door status
Access Badge scans, zone occupancy
ANPR License plates, vehicle metadata
Files CSV, Excel, JSON, SQL dumps
Documents PDF reports, scans, images
Streams MQTT, webhooks, REST APIs

Governance

Evidence-first by design

Intelligence platforms fail without trust. Denerio keeps provenance and access control at the core — not bolted on later.

Immutable raw store

Original files and API responses land in raw/ unchanged. Every downstream record links back to source evidence.

Quarantine, not silence

Validation failures get error codes and owner notification. Data is rejected or repaired — never dropped without a trace.

Lineage & audit

Track who ingested what, when, and how it was transformed. Support compliance and post-incident review.

Sensitivity controls

PII and confidential fields detected early. Encryption, masking, and role-based access before data spreads.

Architecture

Built in layers, deployed in phases

Start with ingestion and validation. Add graph, search, and AI as your data matures.

01

Ingest

Sources & streams

02

Validate

Clean & resolve

03

Graph

Ontology & links

04

Analyze

Search & AI

05

Act

Alerts & workflow

Foundation

Layer 1 — Data source & data lake

Register, preserve, validate, secure, curate, and prepare data before it enters ingestion pipelines. Steps 01–02 above map to this foundation layer.

APIs Databases Files Documents Logs Human input
  1. 1 Source registry Owner, trust level, SLA, access method
  2. 2 Raw data lake Original evidence preserved — CSV, PDF, API, scans
  3. 3 Validate & repair Format, completeness, OCR — reject or quarantine bad data
  4. 4 Dedup & secure Entity resolution, PII detection, encryption & masking
  5. 5 Standardize Normalize dates, units, schema — quality & trust scores
  6. → Graph-ready staging Entities + relationships extracted — hand-off to Layer 2
raw/ Raw lake Unmodified originals
cleaned/ Standardized Validated & normalized
processed/ Graph-ready Entities & links staged

Layer 2 ingestion pipelines load staged data into Postgres, Neo4j, and Qdrant — then layers 03–05 power graph intelligence, search & AI, and alerts. 

# Example: temporal graph query
MATCH (p:Person)-[e:ENTERED]->(z:Zone)
WHERE e.timestamp > $t_start AND e.timestamp < $t_end
RETURN p.name, e.timestamp, z.name
ORDER BY e.timestamp

Rollout

Deployed in phases, not all at once

Start with a working data spine — ingest, validate, query. Add graph, search, and AI as your sources and use cases mature.

  1. Phase 1 · Now

    Event backbone

    Ingest API and CSV → validate → Postgres → operator dashboard. Prove the core loop before adding complexity.

    • FastAPI
    • Postgres
    • React dashboard
  2. Phase 2

    Data lake & graph

    Raw evidence in object storage. Entity enrichment and Neo4j for link analysis and timeline reconstruction.

    • MinIO / S3
    • Neo4j
    • Entity resolution
  3. Phase 3

    Search & AI

    Vector embeddings, hybrid retrieval, and Hugging Face inference through a secure gateway — cloud or on-prem.

    • Qdrant
    • HF gateway
    • RAG
  4. Phase 4

    Act & automate

    Real-time rules, MQTT streams, cases, forecasting, and investigation copilot with full audit trails.

    • MQTT
    • Alerts
    • Workflows

Ready to build on Denerio?

Talk to us about your data sources, security workflows, and deployment needs. We will help you plan the right rollout.

Contact team View architecture