Fake Review Detective

Spot fake reviews before you get burned

Import reviews from a URL or paste them manually. Computes real statistics, then AI scores each review individually and detects manipulation patterns.

Overview

Fake Review Detective uses a two-phase approach: first, JavaScript computes real statistics from your pasted reviews (star distribution, verified %, date clusters, language flags) — instant, no AI needed. Then AI scores each review individually for authenticity (0-100 with red/green flags) and analyzes cross-review patterns (manipulation detection, genuine consensus, purchase recommendation). Every number you see is computed, not hallucinated.

How to use it

1. Paste a product URL to auto-extract reviews, OR paste review text manually
2. Extracted reviews appear in the text area — edit them if needed
3. Select the product category for category-specific benchmarking (auto-detected from URLs)
4. Click 'Detect Fakes' — instant stats appear immediately
5. AI then scores each review individually (Step 1) and analyzes patterns (Step 2)
6. Review the Quick Verdict card for the overall trust score
7. Expand individual review cards to see per-review red/green flags
8. Check the Genuine Consensus section for what real reviews actually say
9. Use the Purchase Recommendation to inform your decision

Example

Scenario: You're considering wireless headphones with 4.5 stars but the reviews seem suspicious — lots of 5-star reviews posted on the same day with generic language, plus a few detailed reviews from verified buyers

What you do: Paste all the reviews, select 'Electronics' category, click Detect Fakes

Result: Instant stats show: 37% verified (red flag), date cluster of 3 reviews within 48 hours. AI scores the generic 5-star reviews at 15-25/100 (likely fake) and the detailed verified reviews at 80+/100 (likely genuine). Quick Verdict: Trust Score 42/100 — 'Approach with Caution.' Genuine consensus: decent sound quality, weak bass, comfortable for short sessions. Verdict: WAIT for more verified reviews.

Tips

• Include as many reviews as possible — pattern detection improves with volume (minimum 100 characters, 3+ reviews recommended)
• Copy reviews with their star ratings and dates for timeline analysis
• The instant stats panel gives useful data even before the AI runs
• Sort reviews 'Most suspicious first' to quickly find the fakes
• The Genuine Consensus section is the most actionable — it tells you what the product actually is based on reviews you can trust
• Try the example reviews to see the tool in action before pasting your own

Common pitfalls

• Don't paste just 1 review — pattern detection requires multiple reviews
• Not every 5-star review is fake — some products are genuinely great
• URL extraction may fail on sites that require JavaScript or block automated requests — paste reviews manually as a fallback
• AI analysis is guidance, not a guarantee — always use your own judgment
• Low trust score means the review SET is unreliable, not that the product is bad