SpecsTool is a free set of tools for the questions you ask yourself before spending money on a PC: are my components balanced? how many frames will I get in the games I care about? is my power supply enough? which graphics card suits my processor? The catalogue currently covers … games with their own performance profile, … processors and … graphics cards.
Looking up how many frames a specific combination of parts will deliver usually ends in a 2019 forum thread, a twelve-minute video that never mentions your resolution, or a calculator that spits out a percentage without explaining where it came from. The information exists, but it is scattered and rarely states the conditions it was measured under.
SpecsTool exists to bring that together into tools that answer quickly and, above all, explain the answer. Every calculator comes with text that says what the number means, what to do about it and when not to trust it. A tool that gives you a figure you cannot interpret has not helped you.
It is an independent project, built and maintained by one person in Spain: a developer who likes hardware and got tired of not finding these answers in one place.
There is no company behind it, no manufacturer, no component store. That cuts both ways. The good part: no vendor pays to score better and there are no affiliate links pushing you toward a particular purchase. The less good part: it runs on personal-project time, so database updates arrive in batches rather than on the launch day of every new graphics card.
Hardware scores are built from in-game performance averages published by independent reviews at 1080p, 1440p and 4K, normalised onto a common scale. Every game in the catalogue has two demand values — GPU and CPU — plus its engine's frame cap where one exists. Game artwork and metadata come from the public RAWG API.
The most played titles are hand-calibrated against published benchmark runs; the rest are estimated from release year, genre and platforms. The methodology page explains the whole process, including the margin of error and what the model cannot know.
The component database is reviewed with each new generation of processors and graphics cards, incorporating figures from the independent testing available at the time. Game profiles are adjusted when better data appears, or when a user sends a correction with their own hardware and real frame rates.
That is in fact the project's best improvement channel: someone who plays a specific title on specific hardware telling us what they actually get.
Every tool is free, there is no sign-up, no paid tier and no data is sold. The cost of the domain, the hosting and the maintenance time is covered by advertising, which is clearly identified and kept separate from the content.
Advertising does not touch the results: the performance model is the same for everyone and no advertiser can influence a score.
It is not a benchmark. It is a statistical model with a typical margin of error of ±10-15%, wider on laptops and for games whose profile is estimated. It is for comparing options and spotting imbalances before you buy, not a guarantee that you will see exactly those frame rates.
It is also not a technical support service or a personalised purchase advisory. If your machine is running badly and you want a real diagnosis, capture frame times while you play: that will show you what is actually happening on your hardware.