1. Input cleanup
The scanner first normalizes what the user submits. A Roblox username and a numeric Roblox ID are treated differently, because typo handling is much more important for names than for IDs. The goal is to avoid wasting scan requests on spaces, formatting mistakes or accidental extra characters.
2. Identity matching
After cleanup, DejaVu tries to resolve the target. This matters because similar-looking usernames can create false confidence. A serious trade check should focus on the correct account, not a nearly identical name.
3. Evidence layers
DejaVu uses available duplicate-risk evidence and public-safe labels. On the website, source names are intentionally simplified into labels such as Local Database and Live Database. That gives traders useful context without exposing internal database names or making abuse easier.
4. Result output
The final report is not meant to replace judgment. It gives a risk signal, evidence count, scan status and readable context. If the result is clean, that means no known hit was found in the available sources at that moment. It does not mean the trade is guaranteed safe.
5. Why reports matter
Community reports help improve the quality of the system when users submit clear evidence. Bad reports are filtered out because low-quality evidence can damage trust. A useful report should include the username or ID, the suspicious item, and why the trade looked risky.