DAI-ATLAS is the software and platform layer — it evaluates, calculates, measures, and projects what is happening on the ground before anyone else knows it.
Flowsint (OSINT graph) + D-Sentinel (multi-agent simulation) built into Atlas — see the graph and cascade without leaving the page.
Jump to graph desk →DAI-ATLAS · MULTI-NODE INTELLIGENCE PIPELINE
The territorial fusion engine observes GDELT headlines and calculates across every node at once — while Flowsint maps the digital entity graph. Atlas outputs a single auditable probability and decision brief.
What: Neo4j graph of domains, orgs, IPs, emails, wallets — enrichers chain WHOIS, DNS, breaches in the same pipeline as field data.
For: Link a mining concession AOI to supplier digital footprint before awarding or responding to incidents.
Output: Entity nodes fed into territorial fusion — not a separate investigation export.
What: Agent mesh ingests GDELT context plus live readings from satellite, VTOL, sensors, and Flowsint — calculates cross-node correlation and territorial stress, not capital-market bars alone.
For: Command rooms that need probability on infrastructure, tailings, borders, and supply chain — from heterogeneous sensors.
Output: Fused node scores → Atlas probability → decision brief with audit trail.
Institutional teams stop juggling tabs. GDELT seeds time context; PeruSAT/Sentinel, VTOL field sorties, terrestrial sensors, and Flowsint enrichers run in parallel. Fusion agents correlate the signals. DAI-ATLAS computes territorial probability and recommends field action — ORBIT BRIDGE within 48 h.
Events stream from all nodes in parallel — same rhythm as a production ORBIT BRIDGE run. Scores drift in real time until live cron replaces demo.
D-Sentinel verifies headlines; GDELT seeds time-stamped context for the AOI.
gdeltSatellite, drones, sensors, and Flowsint entity graph run at the same time — not sequential silos.
flowsintAgent mesh correlates cross-node signals from GDELT + field + digital graph — not market bars alone.
mirofishUnified territorial score — tailings risk, infrastructure stress, escalation likelihood on one tenant.
atlas48 h ORBIT BRIDGE action list — who confirms in field, what to inspect, auditable export.
atlasAll active simultaneously — fusion engine merges signals, Atlas decides.
Temporal seed — verified headlines and geopolitical context
GDELT DOC · ASkDio briefPeruSAT-class + Sentinel change detection on AOI
ORBIT BRIDGE · ESA SentinelField truth — RTK ortho, LiDAR, thermal when program includes OEM layer
CW-15 / CW-25E · mesh relayIoT, seismic, weather, mesh — continuous telemetry into Atlas
SENAMHI · USGS · field gatewaysDigital entities — domains, orgs, suppliers linked to the same AOI
WHOIS · DNS · org→ASN enrichersAtlas territorial probability
GDELT seed: Thermal anomaly at mining concession — GDELT headline correlated with vendor domain
Verified headline · mining beat
NDVI change −12% on NE slope
No recent sortie · schedule CW-25E
SENAMHI precip ↑ · gateway OK
3 vendor entities · exposed subdomain
Recommended action: Deploy VTOL within 48 h · cross Flowsint contractor graph · raise tailings score
5 active nodes · territorial fusion · Atlas score v1
Entity relationships visible now — feeds the same pipeline as GDELT and field layers.
Full workspace when sidecar runs; otherwise live fusion panel from institutional cron below.
npm run flowsint:up · npm run mirofish:ready · GDELT via D-Sentinel cron
HOW DAI-ATLAS WORKS
DAI-ATLAS is not a loose layer map or a market dashboard. It is an institutional pipeline that ingests temporal context (GDELT), observes territory in parallel from satellite, drones, sensors, and digital graph (Flowsint), fuses territorial signals, and delivers an auditable territorial probability plus an ORBIT BRIDGE action list within 48 hours.
Flow: ingest → parallel observation → fusion → probability → decision
Mining teams, civil defense, infrastructure concessionaires, and border operators receive data in silos: headlines in one tab, satellite imagery in another, IoT telemetry on a separate console, digital OSINT in spreadsheets. Nothing correlates across the same AOI with audit trail.
DAI-ATLAS unifies those sources under an institutional tenant. Each run produces a territorial probability score — tailings risk, infrastructure stress, geopolitical escalation, thermal anomaly — and a brief stating who confirms in field, what to inspect, and which assets to deploy (VTOL, mesh relay, satellite window).
After the GDELT seed, four observation nodes run simultaneously. They do not wait on each other: satellite does not block drone; Flowsint does not depend on SENAMHI finishing.
Step 1 — GDELT ingest: D-Sentinel validates headlines and seeds the base scenario with keywords and seed sources.
Step 2 — Parallel observation: each node emits telemetry, metrics, and alerts to the Atlas bus in no fixed order.
Step 3 — Territorial fusion: an agent mesh correlates cross-signals (e.g. GDELT headline + Flowsint subdomain + NDVI drop). The fusion engine does not compute market probability here — it computes multi-source territorial stress.
Step 4 — Atlas probability engine: aggregates per-node scores, applies institutional tenant weights, outputs one percentage with confidence interval.
Step 5 — Decision brief: auditable export with field recommendation, ORBIT BRIDGE 48 h SLA, and custody line.
Flowsint runs as a local sidecar (Docker :5173) or institutional deploy. Enrichers chain OSINT queries on entities tied to the AOI: concession vendors, logistics domains, associated wallets.
The iframe on the Atlas desk shows the live graph; discovered nodes feed territorial fusion as additional signals, not a separate PDF export. Analysts can open the full workspace to pivot manually, but the institutional pipeline already ingested findings automatically.
The fusion engine receives GDELT context plus readings from satellite, VTOL, sensors, and Flowsint entities. Agents simulate correlations and contradictions: if the headline mentions thermal anomaly but satellite shows no change, score drops; if precipitation sensors rise and NDVI falls, tailings score rises.
In production, the fusion sidecar writes rows to Supabase via webhook; the site only reads precomputed results (no visitor blocking). In demo, simulated activity shows the same visual rhythm as a live run.
The fused score is not a simple average: each node has configurable weight by industry (mining prioritizes tailings; infrastructure prioritizes MTC + satellite). Atlas logs which sources participated, timestamps, and model version.
Every output includes auditLine — active node count, live vs demo fusion, score version. This satisfies custody requirements for concessionaires and civil defense.
When score exceeds institutional thresholds, Atlas generates an action list: schedule CW-25E sortie, request PeruSAT window, activate mesh relay in dense canopy, escalate to command room.
ORBIT BRIDGE is the 48-hour commercial SLA from signing: satellite capacity + VTOL deploy + Atlas tenant integration. The intelligence desk you see here is the same logic operating in the client command center.
Demo mode: simulated activity + sample fusion without sidecars. Hybrid: live GDELT from D-Sentinel cron + fusion panel from last Supabase simulation. Full: Flowsint :5173 + fusion sidecar :5174 + cron + webhooks.
Meet the Derteano platform behind predictive territorial control in Peru and Latam.
Learn →Pan and zoom the Andes baseline. Toggle elevation (SRTM). Institutional access adds Sentinel, SENAMHI, MTC, and ACLED layers.
Loading territorial map…
DAI-ATLAS is not a map viewer. It is the intelligent control platform where sovereign data sources become actionable intelligence.
Digital terrain model — slopes, watersheds, and Andean corridors at national scale.
Multispectral change detection — vegetation, mining, deforestation, and urban expansion.
Precipitation, temperature, and alerts — hydrological and mudslide risk in critical basins.
National road network and infrastructure assets — concessioned corridors and megaprojects.
Geopolitical and security events — territorial context for defense and operators.
Authorized OEM VTOL layer (CW-15 / CW-25E) — RTK orthophotos, LiDAR, and video when included in program; satellite intelligence from signing.
Fuse satellite, climate, infrastructure, and events into a unified score per concession or jurisdiction.
Stockpiles, NDVI, water stress, tailings scores, and erosion alerts — auditable pipelines.
RTK 1–3 cm, 2,000 ha per flight — field validation where satellites lack daily resolution.
Claude, DAI-ATLAS, Ollama, and multi-agent swarm — projections before the event occurs.
ArticleHow DAI-ATLAS combines local Ollama, Claude, and multi-agent swarm without losing institutional chain of custody.
Read →
PodcastDAI pipeline: sovereign layers, Atlas v1beta1 API, and B2G deliverables in 14–21 days.
Listen →
VideoMapbox terrain + MapLibre baseline + ops layers — one truth for institutional command.
Watch →Institutional geocoding integrated into the interactive map — Peru first.
More →Lattice-compatible integration for engineering and operations.
More →Public layers and OEM VTOL programs under one territorial control.
More →Access to Command Center, role dashboards, and AI analysis desk.
More →Graph investigations — domains, orgs, breaches — self-hosted Neo4j desk.
View module →Predictive territorial risk scoring — huaicos, tailings, erosion.
View module →Continuous monitoring — ports, highways, transmission lines.
View module →Authorized OEM VTOL (CW-15 / CW-25E) — survey-grade field truth when included in program.
View module →Technical thesis v1.0, Command Center and role dashboards — institutional access.
Enter platform →Predictive Territorial Intelligence · Not a drone company
We unify satellites, drones, IoT sensors, AI agents, and geospatial analysis so governments, mining companies, infrastructure operators, energy companies, and defense agencies make better decisions — faster, with verified data, from any altitude.
DAI-ATLAS is the software and platform layer — it evaluates, calculates, measures, and projects what is happening on the ground before anyone else knows it.
Request institutional access to the full DAI Atlas layer stack and AI analysis desk.
Request demo
