Powered by Omneity Labs

Sovereign AI Models, One API

Build Arabic-first and multilingual workflows with model, function, and knowledge APIs. Choose a task to find its examples and playground.

What do you want to build?

Choose one starting point. Check access before your first request.

Chat & multimodal

Assistants and retrieval

Work with text, images, and audio in Sawalni. Add semantic retrieval with Madmon embeddings.

Speech & transcription

Audio workflows

Transcribe recordings with Daktilo, generate speech with ASAT, and time captions with alignment.

Translation & language

Language workflows

Translate with Tarjamli. Identify languages, transliterate scripts, and encode words with focused APIs.

Knowledge & tools

Search and context

Find news and web sources, look up places, or connect your own documents. Keep source references with results.

OpenAI-compatible chat and embeddings

Use the OpenAI SDK for documented chat completions and embeddings. Speech, translation, language functions, and knowledge APIs use their own documented request formats.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.sawalni.com/v1",
    api_key="your-api-key"
)

response = client.chat.completions.create(
    model="sawalni-beta",
    messages=[{"role": "user",
               "content": "كيفاش نقدر نخدم API؟"}]
)