STARFARER
STARFARER · ENERGY INTELLIGENCE
VYRONYXENERGY INTELLIGENCE · 2026
Safety intelligence · United States · Europe · India

The Next Frontier
of Engineering
Intelligence.

StarFarer is safety intelligence for energy transition assets. Solar, storage and hydrogen fail as one system — their hazards propagate across domains. StarFarer reasons the way your most senior engineers would — across the whole asset, grounded in the US, EU and Indian regulatory stacks and the documented incident record, present at every site, every hour.

VyroNyx — Engineering the energy transition
SCROLL
Solar · BESS · Hydrogen · Process Safety
Live Platform · Enterprise Access

VyroNyx StarFarer

The energy intelligence platform.

Energy organisations run on the judgement of a few senior people — the engineer who has seen this failure mode before, who knows which standard governs, who remembers the near-miss from six years ago. That judgement doesn't scale, doesn't travel, and retires. StarFarer changes that. She answers operational and engineering questions the way your best people would — grounded in encoded standards, the documented lessons of major incidents, and your own organisation's knowledge. Not a chatbot. Not a consultancy. Infrastructure.

Powered by a reasoning graph of standards, statutes and incident knowledge.
Identical inputs. Identical outputs. Every answer sourced and logged.
starfarer · operations console
Reasoning
starfarer ❯ site context — offshore wind + H₂ buffer · North Sea
loaded … 14 assets · 3 jurisdictions · your site standards
"Can we hot-work near the electrolyser hall during tomorrow's turbine lift?"
ASSESS simultaneous operations · H₂ inventory live
GOVERNS ATEX zone boundary at 15 m · your SIMOPS matrix §4.2
PRECEDENT comparable lift, 2019 — dropped-object near-miss
ANSWER Not concurrently — isolate, purge, reclassify the zone first
cites → IEC 60079 · ATEX 2014/34/EU · OSHA 1910.119 · site standard VN-SS-07
✓ grounded · deterministic · logged
Engineering accountabilityEvery answer is sourced, reproducible, and reviewable by your engineers. StarFarer accelerates them. It never replaces them.
Grounded

Every claim cites its source.

API, OSHA, IEC, NFPA, ATEX, Seveso III — and the incident record from Piper Alpha to Buncefield — referenced at every finding, not retrofitted afterwards.

Deterministic

Identical inputs. Identical outputs.

StarFarer is engineered infrastructure, not a stochastic assistant. The same question against the same context returns the same sourced answer — auditable, every time.

Yours

Your organisation's knowledge, encoded.

Your standards, your procedures, your senior engineers' reasoning — captured as structured knowledge StarFarer reasons with. It stays yours, and it stops walking out the door.

Under the surface

An intelligence that compounds.

Most AI tools give the same answer on day one and day five hundred. StarFarer does not. Underneath every response is a reasoning architecture — Aby — that accumulates knowledge from every interaction, surfaces its own gaps, and writes new reasoning chains overnight without being asked.

The gap between what a generic model knows and what Aby knows grows with every question your engineers ask. That gap is not a feature. It is a moat — and it is yours.

Layer 01 · Retrieval

The encoded corpus.

Before Aby constructs an answer, she retrieves the facts that govern it — encoded standards, incident reasoning chains, domain relationships — and reasons within them, not around them.

Layer 02 · Reasoning

Five-phase synthesis.

Every answer passes through structured phases — domain classification, cross-domain grounding, precedent retrieval, synthesis, and quality scoring. The phases are not cosmetic. They are the difference between an answer and a finding.

Layer 03 · Learning

Self-directed growth.

Every night, Aby reviews her own reasoning traces, identifies where her answers were weakest, and generates new knowledge chains to close those gaps. No one instructs her to do this. She does it because it is how she is built.

The compounding effect

Every question your engineers ask is a training signal Aby keeps. Every weak answer she identifies overnight becomes a gap she fills. Every interaction narrows the distance between what she knows and what your best engineer knows. The longer she runs inside your operation, the wider the gap between you and the organisations that don't have her.

StarFarer at work

Four moments. The gap it closes.

Not features. The specific situations where the absence of StarFarer costs something real — time, money, safety margin, or a decision made with less than full knowledge.

Offshore Wind · Hydrogen
03:17 · North Sea platform
The situation

Floating H₂ production platform. Unexpected pressure excursion in the electrolyser package — 340 bar observed against a 310 bar design limit. The procedure covers routine deviations. It does not cover this configuration. The senior engineer is in Aberdeen. Shutdown costs £180,000 per hour. The OIM has fifteen minutes.

Without StarFarer

Full shutdown. £180k/hour. Aberdeen call at 3am. Four hours to restart. The decision made on incomplete information by the most experienced person on shift — who has never seen this exact configuration.

The question asked
"Electrolyser package 340 bar against 310 bar design. IEC 62282 operating envelope, credible causes, safe operating window before mandatory shutdown — do we have one?"
StarFarer

Cause differentiation: membrane compression fault vs downstream isolation valve partial closure vs instrument drift — separated by secondary indicators available on the DCS. Safe operating window: IEC 62282-2 §8.4 allows short-duration exceedance to 1.05× design pressure with specific monitoring conditions. Current reading at 1.097× — outside that window. Recommendation: controlled partial shutdown of affected cell stack, maintain auxiliary systems, avoid full cold shutdown which carries higher restart risk. Comparable precedent: Norsk Hydro 2004 electrolyser overpressure incident, root cause and resolution attached.

IEC 62282-2 · ISO 22734 · NORSOK Z-013 · incident corpus
Process Safety · HAZOP
Monday morning · specialty chemicals facility
The situation

P&IDs arrived Friday afternoon. HAZOP starts Thursday. 24 nodes. Conventionally: the consultant spends the weekend building node descriptions from scratch, arrives Thursday with a blank worksheet, and the team generates rather than challenges. Senior engineer day rates spent on work that doesn't require senior engineering judgement.

The difference

The HAZOP session becomes what it should be — a senior engineer challenging a draft, not producing one. The client gets a more rigorous study in less time. The consultant takes 40% more mandates without diluting quality.

What StarFarer does instead
"Generate full HAZOP node analysis for Node 07 — HCl scrubber inlet. All credible deviations, causes, consequences and existing safeguards. Cross-reference incident corpus."
StarFarer

Complete node analysis: 8 deviations, 23 credible causes, consequence chains including toxic release, overpressure, and pump cavitation. Safeguard assessment against IEC 61511 SIL requirements. Automatic flag: "High flow deviation — comparable configuration implicated in Institute, West Virginia 1985 MIC release. Recommend additional guideword: reverse flow through scrubber during maintenance isolation." That flag — the one the guidewords don't cover — is what the chair's experience is for. StarFarer surfaces it before the session begins.

IEC 61882 · IEC 61511 · CCPS RBPS · incident corpus
Oil & Gas · Management of Change
Tuesday · Gulf of Mexico platform
The situation

A valve substitution. The specified valve is on 14-week lead time. Procurement has found an equivalent — same pressure rating, same nominal bore, different manufacturer. Eight years of experience on the form. Twenty-five years of experience retired in 2021. The MOC question: "Like-for-like replacement?" The answer seems obvious.

Without StarFarer

The form gets signed. The valve gets installed. The incident is not a question of if — it is a question of when. The engineer who retired knew. Nobody who signed the MOC did.

The question asked
"MOC review: substituting Inconel 625 body / PTFE seats with 316SS body / PEEK seats on 6-inch high-pressure sour gas line, 180 bar, 85°C, H₂S 12mol%. Like-for-like or full MOC?"
StarFarer

Not like-for-like. 316SS susceptible to sulphide stress cracking above 3mol% H₂S at this temperature — NACE MR0175 requires higher alloy for this service. PEEK seats show accelerated degradation above 80°C in H₂S service. Full MOC required: materials review, updated P&ID, partial HAZOP rerun for affected node. Flag: this substitution pattern matches the contributing factor in a 2018 North Sea sour gas release — incident reference attached.

NACE MR0175 · API 6D · OSHA 1910.119 · incident corpus
Renewable Energy · Project Finance
Week 6 · financial close · 200MWh BESS · California
The situation

Independent engineer's first information request: 47 questions. The lender is waiting. Every week of delay costs money at project finance cost of capital. The developer's engineering team is stretched across three simultaneous projects. The IE is DNV. They are thorough.

The difference

47 IE questions answered in two days instead of six weeks. At a $180M project with 70% debt at 6.5% cost of capital, six weeks of delay costs approximately $850,000 in financing cost alone. StarFarer paid for itself on question 23.

The question asked
"IE question 23: project was designed to NFPA 855-2023. The 2026 edition is now published. Confirm whether a Hazard Mitigation Analysis is required for this installation and whether the deflagration venting strategy remains acceptable. 40ft containers, LFP, 3.44MWh each, two rows of 29, outdoor, Kern County California."
StarFarer

Yes on both counts — and the second is the exposure. Under the 2026 edition, HMA is the default for this installation: the 2023 stored-energy threshold exemptions are gone, and §4.4 requires the analysis to cover thermal runaway initiation and propagation. Larger issue: NFPA 68 deflagration venting is no longer accepted as the primary explosion control strategy — the design must move to prevention per NFPA 69 or a performance-based alternative backed by installation-level testing. If your UL 9540A data concluded at module level, expect the AHJ to ask for unit-level vent-gas ignition data. Recommend commissioning the HMA now and pre-consulting the Kern County AHJ before the IE response.

NFPA 855:2026 · NFPA 69 · UL 9540A · IEC 62933
The Baseline · Proof before commitment

Your plant. Five days. Then you judge.

No demo environment, no sample data, no imagination required. StarFarer analyses your actual plant and shows you what it found — before you commit to anything.

Step 01 · A conversation

Tell us about your facility.

One call, two to three hours, with a senior engineer. What you operate, what data you hold, where the safety questions keep you up.

Step 02 · Your equipment list

Documents you already have.

A node list from your P&IDs, a fluid inventory, a marked-up plot plan. Our templates make it a two-week job at most — usually far less.

Step 03 · The Baseline

Five working days later.

A HAZOP draft for every node. The map of how failures propagate through your plant. Dispersion and jet-fire physics for your credible scenarios. A clause-level regulatory gap register.

Then you judge. If the Baseline shows you nothing worth knowing, walk away and keep the report.

Your data is deleted on request — confirmed in writing within 48 hours.
Book the Baseline Read the manual first
How it works

Two ways into StarFarer.

Start on the platform, or bring your organisation's knowledge inside it. Both run on the same reasoning core.

The platformFor engineers & teams

General energy intelligence.

The full reasoning platform, ready to use. Ask engineering, operational and regulatory questions across renewables, hydrogen, storage and process safety — answered with sources, at the depth a senior engineer would give.

  • Standards-grounded answers across US and European codes
  • Incident-informed reasoning — the documented lessons of major accidents
  • Study support — HAZOP, HAZID, LOPA/SIL drafting with engineer sign-off
Organisational intelligenceFor energy companies

Your experts, everywhere.

We encode your organisation's knowledge — site standards, procedures, engineering decisions, the reasoning of your named senior experts — into a private intelligence layer only your people can query. StarFarer then answers the way your best engineer would, at every site, on every shift.

  • Private corpus — your knowledge, isolated, never shared or trained on
  • Expert reasoning encoded — judgement captured before it retires
  • Deployed to your operation — control rooms, field teams, night shifts
Where StarFarer works

Built for the energy frontier.

One sector, taken seriously — the places where the energy transition is actually being engineered.

Renewable energy

Solar · Wind · Storage

Utility-scale PV, onshore and offshore wind, and battery storage — design review, interconnection, fire propagation, and the operational questions in between.

IEEE 1547 · UL 9540A
IEC 61724 · NFPA 855
Hydrogen

Production · Storage · Use

Electrolyser packages, compression, storage and offtake — with the hazard profile of hydrogen treated as the discipline it is, not an appendix.

NFPA 2 · ISO 19880 · ISO 22734
IEC 60079 · ATEX 2014/34/EU
Off-grid & microgrids

Islands · Remote · Resilience

Hybrid generation, islanded operation, black-start and dispatch — where the nearest senior engineer is a flight away and the questions can't wait.

IEEE 1547.4 · IEC 62257
IEC 62933
Process safety

The discipline underneath

HAZOP, HAZID, LOPA/SIL, dispersion and consequence reasoning — the safety engineering core that every other capability stands on.

IEC 61882 · IEC 61511
OSHA 1910.119 · CCPS
The regulatory ground

Written into the reasoning, not retrofitted.

StarFarer reasons inside the frameworks your regulators, insurers and lenders hold you to — in the US, across Europe, and in India.

United States

Federal · Codes
  • OSHA 1910.119 · PSM
  • EPA RMP · 40 CFR 68
  • API 510 / 579 / 581
  • API RP 500 / 505
  • NFPA 2 · Hydrogen
  • NFPA 855:2026 · BESS HMA
  • NFPA 69 · Explosion
  • UL 9540A
  • IEEE 1547
  • ASME Sec VIII
  • CCPS RBPS

Europe & UK

Directives · Codes
  • Seveso III · 2012/18/EU
  • COMAH 2015 · UK
  • ATEX 2014/34/EU
  • IEC 61882 · HAZOP
  • IEC 61511 · SIS/SIL
  • IEC 60079
  • ISO 19880 · H₂
  • ISO 22734
  • EU Taxonomy
  • CSRD

India

NGHM · Statutory
  • OISD-GDN-206 · PSM
  • OISD HIRA · HAZOP/QRA
  • PESO SMPV(U) 2016
  • PNGRB ERDMP · GSR 523(E)
  • BIS H₂ Standards
  • Factories Act · MSIHC
  • CTE / CTO · SPCB
  • EIA 2006
And the incident record beneath them all — Piper Alpha · Texas City · Buncefield · Bhopal · Jaipur — encoded as reasoning, so their lessons are applied, not just remembered.
Why trust it

Born in practice, not in a pitch deck.

IChemE Hazards 36Accepted

Peer-reviewed at the frontier.

Paper 162 — propagation containment in integrated solar-BESS-electrolyser architectures — accepted for oral presentation at IChemE Hazards 36. Full paper in peer review, 2026.

IChemEHazards 36Paper 162
Proof of concept

The engineering firm came first.

VyroNyx delivers process safety and energy engineering as a practice. StarFarer was built inside that practice — on real studies, real drawings, real regulatory submissions — before she was offered to anyone else.

HAZOPHAZIDLOPA/SILH₂ Safety Cases
The moat

Expert reasoning, not scraped text.

StarFarer's knowledge is hand-encoded by senior engineers — reasoning chains, not documents. That is slow to build and impossible to shortcut, which is exactly why it works.

Encoded standardsIncident corpusSME reasoning
Contact

Book a conversation.

No sign-up flow, no trial funnel. Tell us about your operation and the knowledge you're trying to scale. A senior engineer replies within two working days.

Operating
United States · Europe
Engineering base · Bengaluru
Or send a message directly