Language AI,
search & MCP

We train translation and transliteration models, build search systems with embeddings, and connect our public library to AI assistants through MCP. These projects have different data and access requirements.

What each system does

Translation changes the language of a passage. Transliteration changes its script. Retrieval finds source material, and MCP gives compatible applications a way to request it.

01

Translation models

Five translation directions across Urdu, Hindi, English, and Arabic, trained on Hadith material prepared by QASRA. The models page records example training runs and explains their limits.

Translation work →
02

Transliteration

Smaller models convert Urdu and Arabic text into Hindi or Roman output. The available bundles run with ONNX Runtime and include tokenizers and inference scripts.

Transliteration details →
03

Embeddings & semantic search

Our Hadith retrieval systems use Voyage and OpenAI embedding models. Separate projects search Quran translations and Tafseer. Retrieved records lead back to the original source text.

How retrieval works →
04

Library MCP server

Compatible AI assistants can search the Maktab-e-Hayat catalogue and retrieve passages with volume and page references. The public catalogue contains 48,000+ books.

MCP setup guides →

Building with QASRA AI?

Tell us whether you need translation, source retrieval, or help integrating an existing tool. For embedding projects and model access, email support@qasra.org with your data, language requirements, and expected usage.

Phone

+91 7737378656 · +91 9529175756

Email

support@qasra.org

Email your enquiry

Describe your project, the material or software involved, and any deadline. For integrations, include your platform and expected usage.

Email support@qasra.org

You can also use the contact form on our home page.