Translate with Tarjama AMT API

Built for Business Language

The quality of an MT engine relies on the quality of data fed into the machine. Translate Tarjama is built using high-quality business data pre-approved by human linguists, resulting in accurate business-centric translations.​

Tailored to Your Data​

Request a tailored Arabic Machine Translation model that’s custom-trained using your own data in the languages you need – resulting in even higher accuracy!​

Maximized Data Security​​

When using Tarjama AMT, your data is protected with the highest security measures in a completely secured cloud environment. We follow the standards of ISO 27001 and guarantee your data is never used or shared with 3rd parties.​​

Integrates with Ease​​​

Tarjama's Arabic Machine Translation supports integration with all the CAT tools you need. Additionally, it protects your data in a secured cloud environment with data and product encryption.​​​

What makes Tarjama AMT different?​

Trained on high quality human translators with decades of experience

Evaluated by human experts

Built for business use only

Adapted to the latest advancements with MT technologies ​

Tarjama AMT continuously outperforms other public MT engines and achieves a higher BLEU score than Google Translate and Amazon Translate.​

How Tarjama AMT works:

Tarjama MT API Description

Service details:

This API uses DRF- Django REST Framework to serve POST requests.

Security Measures:

  • Encrypted traffic using TLS between client and the API
    MT API using JWT for authentication
    MT-API is running in a private subnet, and infront of it a load balancer that forwards the traffic via port 443,80 to the Backend



Body Parameters:

  • languagePair (language from and to, example en-ar)
    accessToken (token to authorize the request)
    translationDomain (default is tarjama-general)
    inputText (Text you wish to translate
    client (Custom MT for a client)

Server response:

Server respond in JSON format as the below:


“originalText”: “Good morning”,
“translatedText”: ” “,
“modelVersion”: “nmt-generic-en_ar-1.7.5”


Why don't you try it yourself?