Kanshi
AI Moderation

Context-aware AI moderation with human review where it matters

Kanshi helps teams moderate comments, posts, reviews, tickets, applications, and listings by routing each submission to allow, reject, or needs review.

Start Free TrialView API guide

Automatic moderation

Clear legitimate and clearly abusive submissions can be handled faster with structured AI decisions.

  • Approve
  • Reject
  • Explain why

Manual review queue

Uncertain submissions stay reviewable instead of forcing a false binary decision.

  • Needs review
  • Confidence score
  • Category

Human-in-the-loop control

Teams can use Kanshi as an assistant that makes moderation faster while humans retain control over edge cases.

  • Reviewable reasoning
  • Operational consistency
  • Less repetitive triage
Decision Engine

Allow, reject, or flag with explainable confidence

Every submission can return a decision, confidence score, category, and reason.

Incoming submission

"I need a quote today. Please call me about your service package."

Confidence
92%
Risk
Low
Category
Valid lead

Allow

Lead intent fits the business and can enter the workflow.

Reject

Submission is fake, low quality, irrelevant, or not worth routing.

Flag

Submission may be real but should be reviewed before entry.

Lead Review Workflow

Automate obvious lead decisions and keep humans in control

Kanshi helps teams allow valid leads, reject fake submissions, and route uncertain cases to review.

Lead review queue

Optional review
FlagQuote request84%Needs review
RejectContact form96%Low-intent pitch
AllowBooking request91%Valid lead
RejectLead form92%Fake inquiry
Use Cases

One validation engine, many submission types

Contact forms are one application. The same context engine can validate reviews, listings, tickets, applications, and community content.

Fake enquiries and quote requests

Identify low-intent contact attempts before they become sales follow-up.

AI-generated spam

Catch polished synthetic content that avoids obvious keyword patterns.

Comment and review moderation

Check whether public contributions are authentic, relevant, and useful.

Marketplace listings and seller applications

Flag scam listings, suspicious vendors, and content that does not fit the marketplace.

Support tickets and feature requests

Separate real customer requests from fake, duplicate, or off-topic submissions.

Registrations and applications

Screen user registrations, vendor applications, job applications, and public forms.

Knowledge base and wiki contributions

Review edits and article suggestions for relevance before they affect public content.

Regulated intake flows

Add context-aware screening to healthcare, finance, insurance, government, and legal forms.

Developer API

Gatekeep leads from any form workflow

Use Kanshi in contact forms, quote requests, bookings, support requests, registrations, and custom lead pipelines.

Forum post

I need help restoring my account access after changing my email.

Product review

The replacement part arrived on time and fit the older model correctly.

Support ticket

Our billing page shows two failed invoices after updating the card.

Marketplace listing

Lightly used commercial mixer with serial number and pickup details.

POST /api/validate
{
  "type": "contact_form",
  "site_context": "Local service business accepting quote requests",
  "content": "I need a quote today. Please call me about your service package."
}

{
  "decision": "allow",
  "confidence": 0.91,
  "category": "valid_lead",
  "reason": "The message matches the business, page intent, and expected buyer request."
}