Fault modeling AI-assisted In your browser

Model your system once. Generate everything.

Draw your system — or describe it and let the built-in AI agent wire the model. FaultTrace derives the FMEA, D-Matrix, fault trees, simulation, and test coverage from that one model. Change the model; everything follows.

Get started Try for 30 days

● Run on an isolated server with yourcompany.faulttrace.app

The FaultTrace editor after auto-clean: a drilled-in Power subsystem with two PSU components, orthogonal auto-routed wires into a Loss-of-power AND gate, a Power BIT test probe on the wire, and the live FMEA docked below — The actual editor — demo project, dark theme. Every screenshot on this page is the real product.

Built for

Teams who can't afford a wrong analysis.

Reliability engineers

Generate the FMEA from the model — not the model from the FMEA.

Test engineers

Prove detection coverage. Justify every test in the suite.

Safety programs

Fault trees with cut sets and numbers you can defend.

Aerospace, defense & automotive

Analyses that hold up through revisions, audits, and recalls.

From your model

Five generators. One source of truth.

Disconnected tools drift. In FaultTrace, every artifact falls out of one model — and every panel below is a real screenshot of it.

FMEA / FMECA

Failure modes, S·O·D, color-banded RPN, MIL-STD-1629A criticality. Every row navigates to its element. Exports to CSV.

D-Matrix mapping faults to the tests that detect them, with ambiguity-group badges and headline coverage metrics

D-Matrix / Testability

Fault × test matrix: % detectable, % isolable, ambiguity groups, undetected faults — the gaps at a glance.

Auto-generated fault tree with OR and AND gates from the propagation model, with what-if fault injection

Fault Trees

Top-down FTA derived from your propagation logic. Click any event to inject it — live what-if.

Simulation view mid-run: playback transport, monitor lanes with a red trip band, the bus voltage waveform, and a trip chip reading Voltage monitor t=610 h

Data Simulation

Seeded time-series runs: shape waveforms, inject faults, scrub the timeline. Reproducible bit-for-bit.

Test-Detection Analysis

Which tests catch which faults, coverage, ambiguity groups — and the redundant tests you can cut.

Headline coverage metrics: Detectable 100% (6/6), Isolable 0% (0/6), Ambiguity groups 2, Undetected 0

Live headline metrics from the D-Matrix view

Built-in AI assistant

Describe the system. Watch the model build itself.

The assistant speaks the whole tool — over a hundred model-building and analysis actions. Every step it takes is auditable, and undoable.

Your key, your control — bring your own AI credentials.
Off by default — no key, no AI egress; admins can disable it instance-wide.
Every action visible — same audited service layer as a human user.

Assistant · building from your description

"Model a redundant cooling loop: two pumps in parallel, a flow sensor, and a temperature monitor that trips on overheat."

✓ create_block Cooling Loop / Pump A, Pump B, Flow Sensor

✓ add_fault Pump A: no flow (λ 2.4e-5/h) · Pump B: no flow

✓ create_gate Loss of cooling — AND(Pump A, Pump B)

✓ attach_test Temp monitor → Loss of cooling

✓ run_analysis cut sets: {Pump A, Pump B} · PoF 5.8e-7

Representative transcript — actions shown are real assistant tools

Try it right here

Inject a failure. Watch it propagate.

This panel is live. Pick a failure mode and watch it propagate, light up the D-Matrix, and recompute coverage in real time.

Try for 30 days

FAULTTRACE ● MODEL

COVERAGE — %

Inject failure

System model

D-Matrix · failures × tests

FMEA · top RPN

—

Detected by

—

Ambiguity group

—

Pick a failure mode above — watch FaultTrace propagate it and show which tests catch it.

Case study · worked example

One model. Five artifacts. Zero drift.

The compressor fault model in the FaultTrace editor: the Vibration test selected, its detection expression shown, and the 33 faults it captures listed in the inspector — The compressor model — the Vibration test and the 33 faults it captures, live on the trace

A compressor drivetrain — motor, bearings, and valves, with vibration and current-monitor tests watching them. Modeled in an afternoon; FaultTrace generated the FMEA, fault tree, and D-Matrix from that one model, then ran 10,000 seeded simulation runs. Twelve faults slipped past the monitors — coverage gaps found weeks before the design review, not during it.

Afternoon to model

10,000

Seeded simulation runs

Gaps caught pre-review

98.2%

Detection coverage

12 coverage gaps caught before design review

Try the live demo

Enterprise-ready

Isolated. Auditable. Yours.

Security & control

Admin / developer / viewer roles enforced server-side. Seats you manage. Instant session revocation. Refuses to boot with default credentials.

Quantitative rigor

Minimal cut sets. Top-event probability three ways. Birnbaum & Fussell-Vesely importance. Deterministic engine, verified against textbook cases.

Your data, no lock-in

Projects export as one JSON document. Print-ready reports. A versioned REST API for everything the UI does.

Dedicated infrastructure

Your own isolated instance at yourcompany.faulttrace.app. Models never share a machine. On-prem for regulated programs.

The in-app Admin Dashboard: Users, Licenses, Requests and Settings tabs, role dropdowns, and a create-user form with role selection — Your instance's admin dashboard — users, roles, seats, and access requests

Explore pricing Talk to the team

Pricing

Your own isolated instance.

Every paid plan runs on a dedicated server at yourcompany.faulttrace.app — your data never shares a machine.

Monthly Annual · save 20%

Engineer

For an individual building and analyzing models.

$49 /mo

1 seat, dedicated instance
FMEA, D-Matrix & Fault Trees
REST API with OpenAPI docs
Up to 10 system models
CSV export & print-ready reports

Pro

For an individual who needs unlimited models.

$79 /mo

Everything in Engineer, plus:

Unlimited system models
Priority queue for support
Add seats — $25/seat/mo

Team

For a team sharing one source of truth.

$199 /mo

Everything in Engineer, plus:

10 seats, larger dedicated instance
Data simulation & test-detection analysis
Role-based access (admin / developer / viewer)
Unlimited models
Add seats anytime — $25/seat/mo

Enterprise

For regulated programs and large fleets.

Everything in Team, plus:

Custom

Talk to us

Custom seat count, sized to you
SSO / SAML & audit logs
Dedicated hardened instance & SLA
On-prem / private-cloud option
Priority support & onboarding

Your dedicated instance is provisioned automatically on subscribe · billed monthly or annually via Stripe · 30-day trial: one-time $5, no auto-renewal — ends after 30 days unless you purchase a plan · add seats or change plans anytime from your in-app admin

Compare plans

See exactly where each plan draws the line.

Plan	Engineer $49/mo Subscribe	Pro $79/mo Subscribe	Team $199/mo Subscribe	Enterprise Custom Talk to us
Modeling & analysis
System models	10	Unlimited	Unlimited	Unlimited
FMEA / FMECA, D-Matrix & fault trees
Data simulation & test-detection analysis	—	—
Built-in AI assistant (your own key)
REST API with OpenAPI docs · CSV export
Team & access
Seats included	1	1	10	Custom
Extra seats — $25/seat/mo	—
Role-based access (admin / developer / viewer)	—	—
Your server
Dedicated isolated instance at yourcompany.faulttrace.app
Instance size	Standard	Standard	Larger	Sized to you
SSO / SAML & audit logs	—	—	—
On-prem / private-cloud option · SLA	—	—	—
Support
Support level	Standard	Priority	Priority	Dedicated + onboarding

Frequently Asked

Questions engineers ask.

What exactly do I build, and what comes out of it?

One fault model: components, failure modes, how faults propagate, and the tests that detect them. From it, FaultTrace generates your FMEA / FMECA, D-Matrix, fault trees, simulation runs, and coverage analysis — always in sync.

Does it really run entirely in the browser?

Yes — FaultTrace runs in any modern browser. Nothing to install or update. Each paid account runs on its own dedicated instance at yourcompany.faulttrace.app.

What does "dedicated instance" mean for my data?

Your account is provisioned onto its own isolated server — your models never share a machine, a database, or a process with another customer's. Spun up automatically when you subscribe, reachable at your own subdomain over HTTPS.

Can I export to the tools and formats my team already uses?

FMEA and D-Matrix export to CSV; every analysis generates a print-ready report. Projects export and re-import as one JSON document you can version-control. And everything the UI does, the versioned REST API does too.

How does billing work?

Billing runs through Stripe — card details never touch our servers. Monthly or annual (save 20%). The 30-day trial is a one-time $5 on a starter-size instance — it never auto-renews and there's nothing to cancel; it simply ends after 30 days unless you purchase a plan, in which case your instance and models are kept and upgraded automatically. Subscribing provisions your dedicated instance automatically — the confirmation page shows progress live and hands you your instance URL and one-time admin login, usually within 2–5 minutes. On Pro and Team, add seats anytime at $25/seat/mo; change plans or buy seats from your instance's admin area via Stripe's billing portal, prorated automatically. Cancel anytime; your instance and data stay available through the end of the paid period.