Recorded sizes, visible sources and explicit limits.
A familiar shoe is a useful starting point when comparing a new pair. RunSized brings together recorded comparisons while preserving model versions, size labels and evidence limits.
A useful comparison describes one person’s experience: “I wear [Reference Shoe] in size [X] (fits me [rating]), and I bought [Target Shoe] in size [Y] which fits [rating].”
Our imported archive still needs review for contributor attribution, exact variants and successful fit. A recorded purchase does not establish which size you should buy.
We also extract shoe-fit facts from published sources and check them against the original report with AI. Each report links to its source. When the same tester reports two shoes, we can show their recorded sizes together, including comparisons linked across separate reports.
These are the tester’s reported sizes, not a prediction for your feet. Missing sizes and widths stay unknown. Several comparisons from one tester still represent one person’s experience. AI checks are not independent human validation.
Record counts describe available material. They are not calibrated accuracy scores or counts of verified independent runners:
See it in action — search for a shoe
Find my sizeThe review checks source statements, same-person comparisons, exact shoe versions and lasts, size systems, widths, duplicate records and fit outcomes. Related versions are labelled separately. We do not turn record counts or AI confidence into a promise of fit.
Clear, first-hand reports help us improve the evidence. If you've bought a running shoe recently, it takes 90 seconds to report how it fits. That data helps the next runner get it right.
Report your fit