Raiplus

Alpha 1.0

API-first content moderation for the Indian internet. We handle both Hinglish and English — the only moderation engine built for code-mixed Indian content.

// the problem we solve

Standard English-only moderation tools miss 60%+ of toxic content on Indian platforms. Our Raiplus Classifier v2 model was trained on 50,000+ Hinglish and English toxic samples — it understands code-mixed insults, leet-speak (c.h.u.t.i.y.a), regional slangs, and common English abuse. Raiplus Refiner v1 then rewrites detected content into professional, human-sounding alternatives.

Node 1 — Guard

Raiplus Classifier v2

Model sizeHigh-Precision Quantized

Training data60,000+ samples

LanguagesHinglish + English

Latency<50ms

Threshold0.85 confidence

HostSecure Inference Node

Node 2 — Sanitizer

Raiplus Refiner v1

Modelraiplus-refiner-v1

Temperature0.6

Latency<400ms

TriggerOnly if toxic = true

HostProprietary AI Cluster

OutputProfessional rewrite

// tech stack

Frontend

Vanilla HTML + CSS + JS

Backend

Node.js + Express

AI Detection

Raiplus Classifier v2 (Secure Inference Node)

AI Rewrite

Raiplus Refiner v1 (High-speed Inference Node)

Auth

JWT + OTP

Database

MongoDB

Rate Limit

In-memory / Redis

Deploy

AWS EC2 + Nginx

// alpha roadmap

✅

Alpha 1.0.4 — current

Raiplus Classifier v2 detection + Raiplus Refiner v1 rewrite + JWT auth + OTP + Rate limiting + Live feed

🔄

Alpha 1.1.0 — upcoming

Private API keys + Dashboard analytics charts + Webhook support

📅

Beta 1.0

Full API access + Team plans + Custom model fine-tuning

Ready to try it?

open playground → read docs → contact us →