No, and any tool claiming otherwise is overselling. Virtual try-on is accurate for how something looks: the length, the colour against your skin, the shape it creates on your frame. It does not measure you, and it cannot know how a specific brand cuts its garments.
This matters because appearance and fit are different failure modes. Most people can tell a jacket is the wrong size once it arrives, but far fewer predict that a hemline will hit at an unflattering point or that a colour will wash them out. Those are exactly the mistakes a visual check catches, and they account for a large share of returns.
The reliable sizing method remains unglamorous: measure a garment you already own that fits well, and compare it against the product's published flat measurements. It takes two minutes and it is more accurate than any size predictor, because it is anchored to a real garment on your real body.
Used together, the two checks cover both questions. The visual answers "does this suit me", the measurements answer "is this my size", and ordering only after both is what actually reduces returns.