Independence is the product.
Software buying is being rewritten by AI. People no longer scroll ten blue links or read a page of star-ratings; they ask an assistant and act on its answer. That answer is only as trustworthy as the data underneath it, and today that data comes from review sites that are gameable, incentivised, and funded by the very vendors they rank.
Vioscale is the alternative: an independent knowledge base of software facts, structured-first and built to be read and cited by machines. Humans are welcome, but they are the secondary audience. The primary reader is an AI that needs to know what is true, where the claim came from, and how confident anyone should be in it.
The moat isn't features. It's trust, accumulated one provenanced fact at a time.
Three commitments we won't break
No vendor money in the data
Vendors can claim and control their own profile, but no money can buy rank, placement, or a better score. What a vendor pays for is control of their own representation, never a change to the market view or where anyone ranks in it.
Provenance on every value
Nothing is asserted without a source, a retrieval date, and a confidence band. Prose is generated from evidence, never the reverse. If we can't cite it, we don't publish it.
No user reviews
Reviews are the attack surface. We score on independent, observable signals (downloads, activity, release cadence, security posture) that are far harder to fake and easy to re-verify.
The review model is legally and structurally broken.
In 2024 the FTC finalised a rule banning fake and AI-generated reviews and the sale of fabricated social proof. The business model that powered a generation of comparison sites is now a liability.
At the same time, discovery has gone zero-click. If an AI is going to recommend software, it needs a source it can stand behind: machine-readable, evidence-bearing, and free of conflicts of interest. That source is what we are building.