Sentinel turns the State Estimator case your control room already has into autonomous, recovery-validated contingency intelligence — at a fidelity no EMS or real-time tool delivers today.
The tools that screen the grid for contingencies were built around tradeoffs the system can no longer afford. Three of them show up every shift.
The industry standard for large-scale contingency screening trades away reactive power and voltage fidelity — exactly the physics that matter most under today’s load growth and IBR penetration.
Even full analysis run against a model snapshot is only as good as the model. The live system’s actual dispatch, topology, and load present a different risk profile than any forecast predicts.
Legacy tools require manually maintained *.con definitions that expire the moment topology changes or equipment returns. Topology changes happen continuously. The lists don’t keep up.
Data centers, EV adoption, and industrial electrification are driving load curves the system was not built to serve. New generation — much of it variable and geographically distributed — is connecting faster than planning models can absorb.
Reliability standards require operators to demonstrate thorough contingency analysis. Regulators increasingly expect operators to show their work — not just that contingencies were screened, but that the screening was comprehensive and the methodology sound.
The limitations of DC approximation are not new knowledge. What has changed is that those limitations now produce consequences visible to operators, planners, and regulators. The question is no longer whether better analysis is desirable — it is whether it is achievable at operational scale.
Scale or physical fidelity — never both. Raptrix Core removes the tradeoff that defined contingency analysis for a generation.
Sentinel and Forge share the same solver, the same Q-limit enforcement, and the same .rpf interchange — so operations and planning can finally work from the same high-fidelity view of the true grid. Studio makes every finding immediately navigable.
Autonomous, recovery-validated N-1 + intelligent N-2 on the live State Estimator case, with auditable certificates.
Same high-fidelity core extended with an integrated LP to position reserves and batteries ahead of expected operating states.
Outage approval, next-day studies, what-if scenarios, and reliability evaluations on the identical physics engine.
The investigation cockpit for any .rpf file. Planned open-source under MPL 2.0 with no model size limits.
Trust is earned where wrong answers have consequences. Raptrix is built for the deliberate timeline this market requires.
This is a long sales cycle by startup standards. It is the right cycle for this market, and Raptrix is building for it.
In a market where wrong answers have consequences, trust comes from transparency — not proprietary lock-in.
All ingestion and export mechanisms — PSSE, PSLF, CIM — are fully open-source on GitHub. Customers, auditors, and regulators can inspect, validate, and extend the integration layer.
Works with all major EMS vendors via standard file push. Every modern EMS already exports the solved SE case on a timer or button press — no custom plugins, no vendor negotiations.
Every .rpf file embeds the full network state, contingency provenance, solver history, and a deterministic fingerprint. NERC auditors can verify how a result was produced.
Forge does not compete with PSS/E, PowerWorld, or other planning environments. It feeds them empirically grounded contingency data they cannot generate themselves.
Sentinel ingests the case your EMS already exports every 5–15 minutes and runs exhaustive, provably filtered N-1 and intelligent N-2 analysis at N-200+ scale. From tens of thousands of “what-ifs” it surfaces the small number of unrecoverable scenarios — each with an auditable certificate.
Every step is automated. Every filter is auditable. Every output is a canonical .rpf your planners can replay.
RAW, PSLF, or native State Estimator exports — no rip-and-replace.
Graph-theory selection augmented by operator-defined guardrails.
Strict Q-limit enforcement and PV-PQ switching. No FDNR shortcuts.
Greedy reserve dispatch validates a real path back to normal.
Every filtering decision is grounded in defensible physics.
One file carries the operating state forward into planning.
Almost everything a utility needs to run both real-time operations and planning studies — in a single fast, reliable, self-describing file. Compact binary, but deliberately built to stay practical for humans.
Built for 5–15 min real-time loops and massive planning studies on laptop-grade hardware. Forward + backward compatible. Tamper-resistant by design.
Self-describing schema embedded in every file. Forward + backward compatible.
Deterministic metadata.fingerprint over the canonical data.
Same input → same fingerprint. Easy to sign or verify with standard tools.
AI data centers and other large, electronically coupled loads behave nothing like traditional demand. Sentinel treats them as first-class citizens — so operators and planners get a defensible, physics-based view of system risk without waiting on EMS vendor roadmaps.
Modeled natively at the bus, not assumed away with a generic load.
Sub-cycle ramp behavior and load reduction schemes carried in .rpf.
First-class in N-1 / intelligent N-2 screening and recovery checks.
PTDF/LODF that’s aware of inverter-based resources, FACTS, and PSTs.
Every operating state Sentinel evaluates ships with the computational-load profile attached — seasonal envelopes, buildout ramps, voltage guardrails, and reduction schemes.
Planning-grade studies for candidate locations and POI/POW decisions — defensible diligence delivered on the same full-AC engine that powers Sentinel and Forge.
Candidate-site screening and full-AC diligence aligned to your approval process.
Interconnection and withdrawal studies at planning-grade fidelity with auditable outputs.
Raptrix began in 2013 as Kestrel when founder Matthew Musto, a power-systems engineer with 20+ years of grid operations experience, set out to solve contingency analysis the right way — full Newton-Raphson physics with no shortcuts. He spent those years inside the control room as Principal Engineer at NYISO until April 3, 2026, when he left to build Raptrix full-time.
This is not a software company that discovered a grid problem. It is a practitioner who understood the problem from the inside — now leading a growing team shipping the only modern platform that pairs a high-performance C++ solver with real-time contingency analysis (Sentinel) and short-term planning (Forge) on one physics engine.
Full-AC Newton-Raphson with strict Q-limit enforcement and intelligent PV-PQ switching — non-negotiable.
Automation that augments operator judgment — never replaces it.
On-prem (RHEL certified), air-gapped, or cloud. No vendor lock-in.
Bring a recent State Estimator export. We will run an exhaustive N-1 + intelligent N-2 with recovery validation and walk through the results with your team — no slide-deck theater.
A pilot engagement on your system — no workflow change, no day-one commitment to act on outputs. Just the chance to see what you have been unable to see.
Our open-source I/O layer creates natural interoperability. We are not trying to displace your platform — we are trying to make it better.
Engagement, validation, and co-authorship of the conversation about what rigorous contingency analysis should look like.
info@raptrixpower.com · Founder direct: matthew.musto@raptrixpower.com · github.com/RaptrixPowerFlow
