Skip to content
AI Studio for Aviation

We design AI for aviation.

An applied-AI studio building real products for aviation — LLM reasoning grounded in real regulation and live flight data. Two products already live, with our flagship safety platform running on Google Cloud.

TWO PRODUCTS LIVE · LLM + RAG ON GOOGLE CLOUD · GROUNDED IN REAL REGULATION

Flagship products

Two products, live.
One AI engine for safer flight.

01

SAMARIA

Live Flagship

An AI-native Safety Management System. It reads every safety report, investigates it with a large language model, and traces each finding back to the exact regulation behind it.

Explore SAMARIA
  • AI investigation Claude (Anthropic) reads each report and returns a structured analysis: severity, root cause (5-Whys / HFACS) and corrective actions
  • Grounded in regulation Retrieval-augmented over 1,647 indexed documents from FAA · ICAO · EASA · Military · NASA ASRS — across four reasoning layers, not just one
  • Traceable, or flagged Every citation must trace to a real source passage; anything it can't trace is flagged as a hallucination, never shown as fact
  • One engine, any actor Multi-tenant — the same engine adapts to airline, flight school, MRO or ATC authority through a profile
  • Live Deployed on Google Cloud (Cloud Run), seeded with 189 real ASN 2025 accident reports
02

Zefyrio

Live

Real-time aviation weather intelligence and a drone-safety HUD that turns a flood of live data into a single GO / CAUTION / NO-GO call.

Try it live — no sign-up
  • Live data fusion METAR/TAF · NOTAM · SIGMET · ADS-B traffic · airspace · radar · GOES satellite — unified in one view
  • Decision engine Resolves wind, gusts, visibility, KP-index and airspace into GO · CAUTION · NO-GO per pilot profile
  • Anywhere Progressive web app + cockpit HUD, usable before and during the flight

How we build

Serious AI
for a regulated sky.

We build AI products for an industry where being wrong has consequences. That means grounded models, reasoning you can trace back to a regulation, and infrastructure built to scale from a single flight school to a national authority — with our flagship engineered on Google Cloud.

  • 01

    AI-native, not AI-added

    Large language models are the core of the product, not a feature bolted on at the end. The reasoning is the value.

  • 02

    Grounded & traceable

    Retrieval-augmented generation over real regulatory corpora. Every citation must trace to a real source passage; anything the model can't trace is flagged, not shown as fact.

  • 03

    Designed around real standards

    We build around the standards the industry trusts (ICAO, FAA, EASA, AS9100) and toward the frameworks governing AI itself: NIST AI RMF and the EU AI Act.

  • 04

    Built on Google Cloud

    Our flagship runs on Cloud Run + Cloud SQL, with a Pinecone vector store and Claude (Anthropic) as the reasoning layer.

0%

Findings traced to a real source

no hallucinated compliance

0

Regulatory documents indexed

FAA · ICAO · EASA · Military · NASA ASRS

0

Accidents fully analyzed by the AI

from 189 real ASN 2025 reports in the system

What's next

One AI engine.
More of the sky.

SAMARIA and Zefyrio run on the same idea: grounded AI applied to a hard aviation problem. Every new problem is a new profile on the engine, not a new company. Here's where it goes.

  • Predictive risk

    From investigating reports after the fact to forecasting where the next incident is most likely — across a fleet or an airspace.

  • Cockpit-native AI

    Hands-free AI briefings and voice reporting on the flight deck, built on the same grounded reasoning engine.

  • Authority scale

    Nation- and regulator-level safety intelligence — the same engine, a bigger jersey.

Beyond aviation

Outside our own products, Nodyt builds custom AI for companies — from small businesses to enterprises: automation, agents and decision systems, engineered to the same standard as everything above.

Get in touch

Let's talk.

Partner, operator, or investor: if you care about safer aviation, we'd like to hear from you.