Free trial
A free forever plan needs no credit card and offers about 15 questions or 3 screen reviews a month, plus 50 free guidelines. A 5-day full refund covers the first paid purchase.
Learn moreCited, pixel-measured UX audits and answers for designers and founders

ProUX is an AI UX assistant that audits screens by measuring real pixels, scores them out of 100 and cites a verified guideline for every finding. It also offers a guideline library, product insights and multi-model comparison.
ProUX is an AI assistant for user experience work, aimed at product designers, UX teams and founders who want design feedback without waiting days for a review. It combines a question-and-answer assistant, a screen auditor and a reading library, all tied to a curated set of UX, UI and conversion guidelines instead of the open-ended memory of a general chatbot. The audit feature accepts an uploaded screen or a pasted URL. Rather than offering opinions, it computes contrast ratios, tap target sizes and text sizes from the actual pixels, then returns a score out of 100 with findings ranked by severity and the arithmetic shown. Each finding states what to change and names the guideline behind it. The assistant works the same way for open questions: answers cite a numbered guideline so the claim can be checked. Projects store product context so it does not need to be re-explained in each chat. Around the core sit a guideline library (50 entries readable without an account, including eCommerce guidance organised by page type and journey stage) and a collection of real-world product design breakdowns that cover business impact and takeaways. Built-in tools rewrite rough prompts into sharper UX queries, run one prompt across several models side by side, and produce a consensus summary. The models used are Claude, Gemini and GPT, selected per task. The product was created by a designer with 19 years of practice and certifications from Nielsen Norman Group and HFI. It suits people who need defensible, source-backed feedback, including non-designers, since answers explain the problem, why it matters and the fix. Compared with asking a general model, the differentiator is the verified library and measured output. Compared with a human review, it trades depth of judgment for speed. A free plan, a paid Pro tier and a five-day refund window make evaluation low-risk, though team seats are not yet offered.


A free forever plan needs no credit card and offers about 15 questions or 3 screen reviews a month, plus 50 free guidelines. A 5-day full refund covers the first paid purchase.
Learn moreA free forever plan includes 100,000 tokens a month (about 15 questions or 3 screen reviews). Pro starts from about $11 per month billed yearly, with 1.9 million tokens (about 215 questions or 70 screen reviews). Quarterly costs more; yearly is 42% cheaper. A 5-day money-back guarantee applies to the first purchase.
Learn moreSeat and team enquiries go to hey@proux.design. An on-site assistant answers questions about the platform, pricing and training, and plans can be cancelled from Settings.
No third-party integrations are listed. The assistant routes work across Claude, Gemini and GPT models depending on the task.
Screen and URL audits scored out of 100 with measured contrast, tap targets and text sizes. Cited UX assistant, project memory, a guideline library, real-world UX insights, prompt optimisation, side-by-side model comparison and consensus summaries.





