KireiFilter
Use case — reviews & ratings

Catch fake reviews
before they skew your ratings.

Fake reviews inflate ratings, mislead customers, and erode trust. Score each review on submission — auto-reject obvious spam, flag suspicious ones for manual moderation, let genuine reviews through.

What KireiFilter catches

Promotional language

Keyword and heuristic patterns catch reviews loaded with promotional copy, excessive exclamation marks, and unnatural phrasing typical of paid or bot-generated reviews.

Suspicious IP addresses

Multiple reviews from the same IP in quick succession, or IPs from known spam sources, are hard-blocked or scored high automatically.

Disposable email accounts

Reviews from throwaway email addresses — a strong signal of fake account activity — score higher, making them easy to filter or flag.

Denylisted senders

Known bad actors in the denylist get a score of 1.0 and are short-circuited immediately — no further processing needed.

Score a review on submission

review.sh
curl -X POST \
https://kireifilter.net/api/v1/spam-check \
-H "Authorization: Bearer <token>" \
-d '{'
"content": "Amazing product, buy now!! Best ever click here",
"ipAddress": "203.0.113.42",
"email": "reviewer@tempmail.io"
'}'
# Response — high score, flag for review
{ "score": 0.78, "verdict": "spam" }

Suggested moderation workflow

1

Score every review on submission. Store the score alongside the review in your database.

2

Score ≥ 0.7 → reject or silently drop. Score 0.3–0.7 → hold in a moderation queue. Score < 0.3 → auto-publish.

3

Human reviewers only spend time on the borderline queue — high-confidence spam never reaches them.

Keep your ratings trustworthy.

Free plan — 100 checks/month. No credit card required.

Create free account