How DoNotStay Works

DoNotStay uses AI to analyze publicly available guest reviews on Booking.com. Our tool reads detailed written reviews — not numerical scores — and identifies recurring patterns related to cleanliness, noise, safety, pests, and other common hotel issues.

Each hotel receives a verdict (Stay, Questionable, or Do Not Stay) and a confidence score. These represent an algorithmic opinion based on review pattern analysis, not a statement of verified fact. Individual experiences may vary.

Red flags are generated when reviews mention notable issues. For critical concerns such as pests, mold, or safety hazards, a single mention is sufficient to trigger a flag. For less severe issues, flags are generated when multiple independent reviews mention the same problem. Each red flag includes the number of mentions and representative quotes from actual guest reviews. Severity levels (Low, Medium, High, Critical) reflect the nature of the issue, frequency of mentions, and recency of complaints.

Our confidence score reflects the absence of serious concerns across reviews. A high score means few or no significant issues were identified. A low score means notable problems were flagged across multiple reviews. The score is not a measure of how positive the reviews are — a hotel with generic but uneventful reviews will score higher than one with enthusiastic praise but recurring complaints about pests or cleanliness.

DoNotStay does not verify the authenticity of individual reviews. We analyze patterns across all available detailed reviews and present our findings as an opinion-based summary.

If you are a hotel owner or manager and believe a verdict is inaccurate, please contact us at disputes@donotstay.app. We take all disputes seriously and will review the analysis.