logo

Developers

The route from a model to a running integration: console, APIs and documentation.

Where you build

The platform products a developer works in, the ones with a live surface first. Keys, the API reference and usage all live in Kawn Console; this page does not reproduce it, because one reference kept in one place is the only version that stays current.

  • Kawn Console

    One account, the whole Arabic model suite

    dashboard.kawn.ai

  • Qurani.ai

    One clean API for Quranic text, audio and translation

    qurani.ai

  • Nova Star

    No-code automation that speaks Arabic and government

  • Seamless API

    Build Arabic AI pipelines visually, not from raw APIs

Models you can call

Every model in the catalogue that records a cloud deployment option, with what it does and where its page is. The catalogue is the source: a model that is not marked for cloud access is not listed here.

  • Kuwain

    A 1.5B-parameter Arabic SLM built by language injection, the base layer for our downstream models.

    Language1.5B

  • Lahjawi

    A family of Arabic dialect translation models covering dialect-to-dialect and dialect-to-MSA translation.

    Language

  • Mutarjim

    Bidirectional Arabic–English translation built on Kuwain, for general and specialized domains.

    Language1.5B

  • Sadeed

    Arabic diacritization built on Kuwain, restoring full Tashkeel for MSA and classical Arabic.

    Specialized NLP1.5B

  • Kawn Embed Light

    Lightweight general-purpose Arabic embeddings for semantic search, retrieval, clustering, and RAG.

    Embedding300M

  • Kawn Embed Islamic

    Arabic embeddings optimized for retrieval across the Quran, Hadith, Fiqh, and Fatwa collections.

    Embedding139M

  • Kawn Embed Medical

    Domain-specific Arabic embeddings for healthcare applications.

    Embedding139M

  • Nabr

    Acoustic model for Quranic recitation analysis, covering ASR and Tajweed-aware mispronunciation detection.

    Speech

  • Sawt Najd

    High-quality Najdi Arabic text-to-speech for conversational AI and accessibility applications.

    Speech

  • Islamic QA Classifier

    Hierarchical classification that routes Fatwa and Fiqh questions to the correct legal category.

    Specialized NLP

  • Islamic Quotation Extractor

    Information extraction that identifies verbatim Quran, Hadith, and book quotations in classical and Fatwa texts.

    Specialized NLP

  • Baseer

    Arabic document intelligence: scans to structured Markdown

    Vision-Language3B

  • Baseer Nakba

    Handwriting-specialised variant of Baseer

    Vision-Language3B

Client libraries

The client libraries Misraj publishes, with the install command each repository's own README gives.

  • kawn.ai Python SDK

    The official Python client for the kawn.ai Models API. It exposes an embedding service and a Baseer OCR service, each with a synchronous and an asynchronous form, and reads its key from the MISRAJ_API_KEY environment variable.

    PythonMIT

    pip install kawn.ai
  • LlamaIndex Kawn integration

    LlamaIndex wrappers for the Kawn SDK: KawnEmbedding for document and query embeddings, and BaseerReader, which reads PDFs and images through the Baseer OCR service into LlamaIndex documents.

    PythonMIT

    pip install llama-index-kawn

Deployment options

Each model records how it can be run. These are the options the catalogue actually carries, and the models under each are the ones that carry it.

Open weights and code

What can be downloaded and read rather than called, and where it is published.

  • Hugging Face

    huggingface.co

  • GitHub

    github.com

  • Open Source

    Models, tools and corpora Misraj publishes openly, and where to find them.

Not on this page

No rate limits, pricing, endpoint list or uptime figures. This site has no source for them and will not guess; the console has all four, and it is the copy that is kept current.

Let's talk about what you're trying to build.

Tell us the problem. We'll tell you honestly whether AI is the right answer.