Manifesto
Athar exists for one reason: to bring the power of modern AI to Arabic and Middle East legal practice without importing its most dangerous habit — confidently inventing the law.
Ask a general-purpose chatbot a legal question and it will answer with total confidence — and a citation that looks perfect and doesn't exist. It guesses article numbers, invents judgments, and dresses speculation as authority. That is a tolerable flaw for a poem. It is unacceptable for the law.
When a lawyer relies on an authority that isn't real, the consequences land on a client, a case, a livelihood. The cost of a hallucination in law is not a wrong answer; it is a sanction, a lost matter, a breach of duty. An AI for legal work cannot treat being wrong as a rounding error.
The tools getting the most attention are built for English-language, common-law practice. Arabic legal language, the civil codes of the region, its courts of cassation and specialized tribunals — these are an afterthought, if they appear at all. Practitioners across the Middle East deserve more than a translation layer over someone else's law.
Athar retrieves before it answers, cites only what it retrieved, verifies every citation against the source corpus, and drops anything that doesn't resolve. When the law doesn't support an answer, it says so. This isn't a setting you enable — it is the foundation the product is built on.
“We would rather Athar tell you it doesn't know than tell you something that isn't true.”
What we believe
Try Athar and see what grounded legal AI feels like — every answer, with the receipts.