Upload failure data — including suspensions — and get a fitted Weibull with probability plot, confidence bounds, and B-lives in seconds. No install, no licence server, free to start.
Most real maintenance datasets are dominated by items that haven't failed: preventive replacements, items still in service, units retired for other reasons. Ignore those suspensions and every life estimate comes out biased low. Reliafy fits by maximum likelihood with full support for right-censored, left-censored, interval-censored, and truncated observations — so the units that survived count as the evidence they are. (If you've been doing Weibull analysis in Excel, this is the part Excel can't do.)
Fit Weibull, Lognormal, Exponential, Gamma, and Normal distributions to the same data — or rank them side by side by AIC with one click and let the data pick. Proportional-hazards models handle covariates when operating conditions differ across the fleet. Every fit reports its parameters, goodness of fit, B10 life, and a probability plot you can read at a glance.
Point estimates aren't decisions. Every fitted model carries confidence bounds on the probability plot and through the survival calculator, so "the B10 life is 4,100 hours" comes with the uncertainty attached — which is what your maintenance interval actually hinges on.
Import a CSV exported from your CMMS or a plain spreadsheet: map the time column, mark which events are failures and which are suspensions, and fit. Datasets are saved and reusable, so next quarter's update is an append, not a rebuild. Fitted models flow straight into the rest of the platform — optimal replacement intervals, reliability block diagrams, and evidence-linked RCM studies.
The cloud free tier includes saved datasets and models with sample data to learn on. The full engine is AGPL-licensed and self-hostable with one command if your data can't leave site. Statistical core built on the open-source surpyval library.
Upload a CSV, mark the suspensions, and read the plot — free.
Get started free