ReviewLabAI

AI review analysis for e-commerce — multilingual, dialect-aware, one script tag.

Drop one <script> tag. Shoppers get an AI chat widget that answers questions from real reviews. You get sentiment, trends, and summaries — across any language mix, automatically.

→ Try the live demo

---

Bring Your Own Key (BYOK)

Most review AI tools lock you into their model and billing. ReviewLabAI lets you plug in your own OpenAI (or compatible) API key — you control which model runs, what it costs, and where the data goes.

# Set your key in the dashboard, then the API uses it for all inference
curl -X POST https://reviewlabai.com/api/settings/ai-key \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "provider": "openai", "apiKey": "sk-..." }'

What this means in practice:

• Use gpt-4o, gpt-4o-mini, or any OpenAI-compatible endpoint
• Your key, your usage dashboard, your cost visibility
• No per-summary markup — you pay OpenAI directly
• Swap providers without touching your integration

→ Full BYOK explainer

---

What it does

• Summarises reviews into pros/cons, sentiment score, and key themes
• Handles any language — Arabic dialects, Arabizi, French, mixed-language, transliterated text
• Embeddable widget — shoppers ask questions, AI answers from review content
• Tag extraction — automatically surfaces recurring topics (shipping, scent, quality…)
• Trend detection — flags when sentiment on a topic shifts week-over-week
• Multi-platform — Salla, WooCommerce, Zid; REST API for anything else

---

Widget (one line)

<script
  src="https://reviewlabai.com/api/widget.js"
  data-merchant-id="YOUR_MERCHANT_ID"
  data-product-id="YOUR_PRODUCT_ID"
></script>

Loads in a Shadow DOM — no CSS conflicts, no framework dependency, ~14 KB gzipped.

---

REST API

Base URL: https://reviewlabai.com/api

Submit reviews

curl -X POST https://reviewlabai.com/api/reviews \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "product_id": "prod-123",
    "reviews": [
      { "text": "ماشاء الله المنتج فوق الوصف والتوصيل سريع", "lang": "ar", "rating": 5 },
      { "text": "Lasts all day, great projection", "lang": "en", "rating": 5 },
      { "text": "3ajeeeb el 3etr, ystahel el price", "lang": "ar", "rating": 5 }
    ]
  }'

The third review is Arabizi (Arabic written in Latin script) — detected and processed automatically, no pre-processing required.

Get AI summary

curl https://reviewlabai.com/api/products/prod-123/summary \
  -H "Authorization: Bearer YOUR_TOKEN"

{
  "sentiment_score": 89,
  "pros": ["Long-lasting", "Strong projection", "Good value"],
  "cons": ["Price point"],
  "top_tags": ["longevity", "scent", "packaging"],
  "languages": { "ar": 4, "en": 2 },
  "summary_en": "Premium fragrance with excellent reviews, stands out for longevity and projection.",
  "summary_ar": "عطر فاخر بتقييمات ممتازة، يتميز بثباته الطويل وقوة انتشاره."
}

Ask a question (the widget call)

curl -X POST https://reviewlabai.com/api/chat \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "product_id": "prod-123",
    "question": "Does this fragrance last through the day?",
    "lang": "en"
  }'

{
  "answer": "Based on 6 reviews: yes — reviewers consistently mention 8–12 hour longevity on skin.",
  "sources": 6
}

---

Multilingual pipeline

Reviews from e-commerce stores often arrive in messy, real-world formats:

Input type	Example	Handled?
Standard Arabic	"المنتج ممتاز والتوصيل سريع"	✅
Gulf dialect	"وايد زين وثمنه مناسب"	✅
Arabizi	"3ajeeeb, ystahel el price"	✅
Mixed-language	"الطلب جاء fast جداً"	✅
French	"Excellent produit, livraison rapide"	✅
English	"Great quality, fast shipping"	✅

The pipeline normalises all of the above before summarisation — no pre-processing needed on your end.

---

Stack

Layer	Tech
API	Node.js · Express 5 · TypeScript
AI	OpenAI-compatible (BYOK) · GPT-4o-mini default
DB	PostgreSQL · Drizzle ORM
Auth	Clerk + HMAC bearer tokens
Widget	Vanilla JS · Shadow DOM · esbuild · ~14 KB
Dashboard	React 18 · Vite · Tailwind · shadcn/ui

---

Platform integrations

WooCommerce

# Install plugin from your dashboard:
wp plugin install reviewlabai.zip --activate
# Reviews sync automatically from that point

Salla

One-click install from the Salla App Store — no code required.

Zid

Point your Zid review webhook at POST /api/webhooks/zid.

Anything else

Any platform that can POST JSON to an endpoint works. The REST API is the integration.

---

Roadmap

• Shopify integration
• Magento / Adobe Commerce module
• AI-drafted review replies (in your brand voice)
• Competitor review benchmarking
• npm package — @reviewlabai/widget
• OpenAPI spec + Postman collection

---

Questions / feedback

Open a Discussion — integration questions, feature ideas, anything.

Direct contact: support@reviewlabai.com

---

Get access

API keys and merchant accounts at reviewlabai.com — free to start, no card required.