Skip to content
Platform

From Collection to Finished Product

Raw OSINT ingestion to IC-standard intelligence products — SITREPs, WARNINTs, and threat assessments produced autonomously. No analyst assembly required.

Demonstrator

Not a synthetic demo. A live intelligence environment.

Ishtar's entry demonstrator covers Iranian conflict dynamics and Gulf energy infrastructure — the intelligence problem that matters most to the defense primes and energy operators we serve. Every evaluation runs against real collection infrastructure, real source data, and real agents producing real products.

portal.ishtar-ai.com/dashboard
OperationsDashboard1 active warning
Warning intelligence activeISHTAR-WARN-003

IRGCN fast boat surge at Strait of Hormuz approach — Bandar Abbas

Active Warnings
1
+1 since 03:44 UTC
Events · 24 hr
347
12 P1/P2 EEI hits
Kharg Vessels
2
↓ below 30-day avg
Sources Active
13/15
2 degraded
Live Event StreamP1 / P2 EEI
F6

IRGCN fast boat deployment reported near Qeshm Island — 8 vessels

04:01 UTCEEI-2.1IRGCN
B2

TankerTrackers: MV Pacific Serenity departs Kharg northbound

03:47 UTCEEI-1.1EEI-1.2
A1

UKMTO Advisory: Suspicious approach in Gulf of Oman

03:30 UTCEEI-2.4IRGCN
B2

AIS gap detected — VLCC Sinopec Atlas (AIS-off 47 hr)

02:15 UTCEEI-4.4Dark fleet
Tripwire StatusSCR-004
T-1Zero tanker traffic 5+ daysNormal
T-2Damage at Kharg infraNormal
T-3IRGCN tanker seizureWatch
T-4Hormuz closure threatNormal
Recent Products
WARN

IRGCN Naval Surge

WARN-003 · 08 Apr

SITREP

Kharg Weekly Export

SITREP-018 · 07 Apr

TA

Houthi Maritime Threat

TA-009 · 05 Apr

From raw OSINT to finished intelligence — in minutes, not days.

Areas of Interest — Entry Demonstrator

Kharg Island
Bandar Abbas
Strait of Hormuz
Natanz
Fordow
Bushehr
Ras Tanura
Abqaiq
Kharg Island SPM
Bab el-Mandeb
Red Sea Corridor
Beirut Dahiyeh
Platform Architecture

Four layers. One finished product.

Every layer of Ishtar is designed to reduce the distance between raw collection and actionable intelligence. The platform does not present data. It produces assessments.

01Tiamat

Data Fusion Engine

Ingests and normalizes data from 15+ OSINT sources — AIS feeds, satellite imagery, GDELT, RSS, social media, and government publications. Every source is tagged with collection metadata and routed to the scoring layer.

GDELT · AISSatellite imageryRSS/ATOMSocial mediaGov publications
02Nabu

Analytical Agent Layer

Six purpose-built AI agents run in parallel — scoring sources against the Admiralty Code, generating competing hypotheses, detecting patterns of life, red-teaming assessments, modeling financial impact, and synthesizing finished products.

Admiralty scoringACHPattern of lifeRed teamSynthesis
03Marduk

Production & Dissemination

Transforms analytical output into IC-standard intelligence products — IIRs, SITREPs, WARNINTs, and threat assessments. Products are formatted to ICD 203, JIPOE, and NATO STANAG standards with full source attribution.

IIRSITREPWARNINTICD 203NATO STANAG
04Enki

Monitoring & Warning

Autonomous tripwire system monitors defined indicators around the clock. When thresholds are crossed, Enki triggers the full agent pipeline to produce warning intelligence — no analyst intervention required to initiate.

Tripwire monitoringAutonomous WARNINTThreshold alertingEEI tracking
The Nabu Agent Team

Six agents. Continuous production.

Each agent is purpose-built for a specific analytical function. They run in parallel, in coordination, around the clock.

Admiralty Scoring

Haiku 4.5

Evaluates every source against the NATO Admiralty Code — reliability of source (A–F) and credibility of information (1–6). Scores propagate through the entire analytical chain.

ACH Agent

Opus 4.6

Generates competing hypotheses using Analysis of Competing Hypotheses methodology. Weighs evidence for and against each hypothesis, identifies diagnostics, and flags assumptions.

Pattern of Life

Sonnet 4.6

Detects deviations from established behavioral baselines — vessel movements, facility activity, communications patterns. Anomalies are flagged with confidence intervals.

Red Team

Opus 4.6

Challenges every assessment before it reaches production. Tests analytical conclusions against alternative explanations, identifies cognitive biases, and stress-tests key judgments.

Financial Impact

Sonnet 4.6

Translates geopolitical events into financial and operational impact estimates — supply disruption costs, insurance rate changes, commodity price effects, and infrastructure exposure.

Synthesis

Opus 4.6

Integrates outputs from all agents into finished intelligence products. Resolves conflicting assessments, assigns overall confidence levels, and formats products to IC standards.

Why Ishtar

Not a better Palantir. A different answer.

The question Palantir answers is: how do we give analysts better tools to explore data? The question Ishtar answers is: how do we produce finished intelligence without requiring analysts to build it from scratch?

CapabilityPalantir GothamIshtar AI
Design philosophyAnalyst-firstAgent-first
Primary outputDashboards & data explorationFinished intelligence products (IIRs, SITREPs, WARNINTs)
Source evaluationUser-applied metadataAutomated Admiralty Code scoring on every source
Deployment modelEnterprise on-prem or FedRAMP cloudCloud-native SaaS, deployable to air-gapped enclaves
Warning intelligenceManual tripwire configurationAutonomous tripwire monitoring with agent-generated WARNINTs
Analyst burdenHigh — analysts build products from raw dataLow — agents produce drafts, analysts review and release
IC product formatsCustom exportsNative ICD 203 / JIPOE / NATO STANAG formats
Time to first valueMonths (integration + training)Days (SaaS onboarding + demonstrator)
Technology Stack

Built to the standard your environment demands.

Ishtar runs cloud-native on AWS with a security posture designed for defense and intelligence deployment from the first line of code.

Compute & Orchestration

  • AWS ECS Fargate
  • AWS Lambda
  • AWS Step Functions
  • EventBridge

Storage & Search

  • Amazon S3
  • OpenSearch
  • DynamoDB
  • ElastiCache Redis

AI / Model Layer

  • Amazon Bedrock
  • Claude Opus 4.6
  • Claude Sonnet 4.6
  • Claude Haiku 4.5

Integration Layer

  • GDELT API
  • MarineTraffic AIS
  • Planet Labs Imagery
  • RSS/ATOM Feeds

Security & Identity

  • AWS IAM
  • Cognito
  • KMS Encryption
  • VPC Isolation

Delivery & Edge

  • AWS App Runner
  • CloudFront CDN
  • Route 53
  • ACM TLS
Use Cases

Built for the environments where intelligence decisions matter.

Defense Primes & J2

Finished intelligence for units that cannot staff a full analytical shop

Program managers and J2 shops need intelligence products, not data exploration tools. Ishtar delivers finished assessments — SITREPs, WARNINTs, threat assessments — formatted to IC standards and ready for operational use.

SBIR/STTR, OTA prototyping, defense prime subcontracting

Learn more →
Energy & Critical Infrastructure

Geopolitical risk expressed in barrels and dollars, not threat levels

Energy operators need risk intelligence that maps directly to operational decisions — cargo routing, facility security posture, insurance exposure. Ishtar translates geopolitical events into financial and operational impact.

Gulf energy demonstrator covers 12 AOIs with daily imagery

Learn more →
Allied Defense Ministries

Sovereign intelligence capability without sovereign cost

Allied nations need independent analytical capability without building a full intelligence infrastructure. Ishtar provides a deployable platform that produces IC-standard products from sovereign-controlled sources.

NATO NCIA, UK DSTL, Canadian IDEaS, AUKUS Pillar II

Learn more →
Get Started

See the demonstrator live.

The Iranian conflict dynamics demonstrator runs against real collection infrastructure. Request a session and your evaluation team sees finished intelligence products produced in real time from real sources — not a scripted walkthrough of synthetic data.

NDA available upon request · CAGE/DUNS registration accepted