Please rotate your device back to portrait
A Compound AI Platform
It checks your model the way a real audit team would - a fast first pass, then careful checks, then a specialist review - all in one run.
Like having a senior model reviewer on call, any time, in minutes instead of weeks.
Traditional model reviews are slow and expensive because they depend on a specialist's time. We built a pipeline that applies the same rigour - real formula-level recalculation, deep domain expertise for the industry your model is actually in - without the multi-week wait or the five-figure invoice.
Simply, upload your financial model and get back a 16-tab structured report - (Issue Log, Validation Matrix, Remediation, Scope and Reliance, and more) that checks the model the way a specialist reviewer would: tracing formula logic, recalculating outputs, and reasoning about whether the numbers actually mean what they claim to mean, not just whether the sheet looks tidy.
For models that fit within a single pass, every formula gets read in full - not sampled. Catches the errors a values-only review structurally can't: a discount-rate mismatch, a metric wired to the wrong line, logic that's technically valid but economically wrong.
The review adapts to what you upload. Mining models get checked against cut-off grade economics and royalty structures. Property models against development and GDV logic. Infrastructure against concession and project-finance mechanics - automatically, no manual setup.
Not a bug list. A structured, multi-tab report - issue log, validation matrix, remediation plan, scope and reliance - the same shape a client-facing review deliverable actually takes.
Every check maps back to an established methodology - the FAST Standard and ICAEW's Financial Modelling Code - not an ad-hoc list of gut checks.
Upload a model and check on a run from a mobile browser - no desktop required to kick off a review or see where it stands.
No need to keep a tab open. The report lands in your client's email the moment it's ready, so a review can run in the background of your day.
Runs on dedicated, root-controlled infrastructure - never shared hosting. Each run executes in its own isolated, purpose-provisioned environment.
Every connection runs over TLS end to end. Uploaded models are encrypted at rest with AES-256, not stored as plain files on disk.
Infrastructure is hosted in-region, so your model never has to leave the jurisdiction your business operates in.
Your model is used to generate your review - nothing else. It is never used to train PlsFx's models or any third party's.
Uploaded files and working data are cleared automatically on a retention sweep - nothing lingers by default once your review is done.
Billing runs through a dedicated payment processor. Card details are never handled or stored directly by PlsFx.
You pay for what the model actually needs reviewed - not a flat subscription regardless of size. A short, simple model costs less than a sprawling, formula-dense one, because the review effort genuinely differs between them.
The f-score reflects formula complexity, not just formula count - so two models with the same number of formulas can price differently if one is genuinely harder to verify than the other. It's the same principle a specialist would use to scope an engagement, applied automatically.
© Copyright 2026 BIRDHOUSE TRADING PTY LTD. - All Rights Reserved / ABN: 61692125616
Financial model review hasn't kept pace with what's actually possible.
Every major decision - a loan, an investment, a board sign-off - sits on top of a spreadsheet nobody outside the room who built it has genuinely checked, because doing it properly has always been slow, expensive, and dependent on one specialist's calendar.
That gap exists for two reasons.
First, most tools that claim to "check" a model only look at what it outputs, not how it gets there. A values-only glance can't catch a discount rate wired to the wrong cell, or a metric that's technically correct but economically meaningless - catching that means tracing the formula logic itself.
Second, a review is only as good as the standard behind it. A mining model and a property development model don't fail in the same places - generic checklists miss what's specific to each.
PlsFx closes both gaps at once.
We recalculate the model's actual formula logic, not just its outputs, catching what a surface-level pass structurally can't see - grounded in a domain-specific methodology built on standards like the FAST Standard and ICAEW's Financial Modelling Code, not gut-feel checks.
The result is a structured, multi-tab report - the same shape a client-facing deliverable actually takes - in minutes instead of weeks, at a fraction of the cost.
Where we're going. This is the beginning of what automated model review can look like. We're expanding industry coverage, deepening how much of a model we can verify in one pass, and building toward the kind of institutional-grade review today's process was never built to support.
Our goal is to be the standard financial teams reach for by default - not a workaround for when a specialist isn't available, but a genuinely better way to know a model is right.