Islamic Embeddings

Overview

The Kawn Embeddings API converts text into numerical vectors that represent its semantic meaning. It provides a single, streamlined endpoint for generating embeddings.

The flagship model — tbyaan/islamic-embedding-tbyaan-v1 — is fine-tuned on Arabic Islamic data including Hadiths, Fatawy (legal rulings), and classical Islamic books. It understands nuanced Sharia and Fiqh terminology, making it ideal for:

• Semantic Search — Find documents that mean the same thing as a query even without shared keywords.
• RAG (Retrieval Augmented Generation) — Retrieve relevant context for LLMs.
• Clustering — Group similar texts by meaning.

Endpoint

POST /v1/embeddings

Generate embeddings for text input. This endpoint is OpenAI-compatible — it uses the same request and response format as the OpenAI Embeddings API, making it easy to integrate with existing tooling built for OpenAI.

Requires an API Key passed via the x-api-key header.

Request Body

Example Request

curl -X POST https://api-dev.kawn.io/v1/embeddings \
  -H "x-api-key: <YOUR_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tbyaan/islamic-embedding-tbyaan-v1",
    "input": "ما حكم الصلاة في الأوقات المنهي عنها؟"
  }'

Models

Response

{
  "data": {
    "object": "embedding",
    "index": 0,
    "embedding": [0.021, -0.003, 0.118, "..."]
  },
  "model": "tbyaan/islamic-embedding-tbyaan-v1",
  "usage": {
    "promptTokens": 12,
    "totalTokens": 12
  }
}