LogicReader is a schedule-intelligence platform for engineering and infrastructure projects. Compare revisions, build defensible logic, assess risk, summarise huge programmes, and audit every dependency — on top of the Primavera P6, MS Project, or Asta schedules you already produce.
Free for as long as you use it and give us feedback — open a schedule, read it, and work on it. We ask for your email so we can send you a confirmation link; nothing is shared. One product, three ways to run it: run it here for free, unlock P6 export, collaboration and bigger projects with LR Plus, or keep it inside your own perimeter with the desktop and on-premise delivery. Web access is what is available today; the desktop build follows — see the three ways in.
Seven things, all on the same canvas. Pick any one as the entry point — the others come along for free. Every box opens its training.
Compare
Two revisions, one canvas
Build
Draw logic from scratch
Assess
Logic health, scored
View
25k tasks, interactive
Summarise
From thousands to dozens
Control risk
Risks on the graph
Audit
Every dependency, traced
Click a box to open the training.
Most scheduling tools are built for one user — the planner. We work so that the engineer, the project manager, the site lead, the commercial lead, the director and the bid manager all get their answer without going through the planner. Not for a procurement portal. For the engineer who actually has to build the thing.
Click a role to open its page in the Library.
A schedule is not a bar chart. It is the moment a looping process — do, check, redo — agrees to stop: every loop unrolled into a fixed number of passes, priced into the budget, signed by everyone. The logic is what gives the dates their meaning.
LogicReader puts that logic on the canvas — visible, interrogable, defensible — so every stakeholder can see why a date is what it is, and stand behind it.
The full argument lives at how we think.
Three capabilities live on the platform right now. No mock-ups.
What makes a schedule clear to a human — this link exists for this reason, decided by this person, at this cost to change — is exactly what makes it computable by a machine. Readable and AI-usable are one property, not a trade-off.
Others teach AI to guess what happened. LogicReader records why it happened — so AI never has to guess. The typed, justified record your team builds every day becomes the ground your own AI, under your control, can stand on.
One principle shapes every product decision. We will trade features for it. We will not trade it for features.
If using LogicReader doesn’t feel better than working in P6, none of the analytics matter. Visual clarity, instant feedback, and graceful behaviour at scale are not features — they are the product.
Why we build it this way — the company principles, the work ethics and the dream — is at how we work.
Data & Trust
Six absolute commitments. No marketing caveats.
Your schedules stay your IP. Export or delete any time.
Behind authentication. TLS in transit. Encryption at rest.
Ever. Not on aggregated data. Not on “anonymised” data.
Not sold, licensed, distributed, syndicated. Period.
No industry reports drawn from your projects. No derived analytics.
Enterprise tier runs entirely in your perimeter. No data leaves.
Where we’re going
CosmosPM is a British company on UK-hosted infrastructure. Our AI provider today is US-based (Anthropic) and used only when you activate an AI feature inside LogicReader — never on stored schedules. As European frontier models (Mistral, Aleph Alpha, and the UK’s emerging AI ecosystem) reach capability parity, LogicReader will offer a UK/EU-sovereign AI option with the same six commitments above carried unchanged. The AI layer is pluggable by design, not hard-wired. No permanent lock-in to US compute. See the sovereignty path →