Legal AI built for India, grounded in primary sources

Why we built Niyam

Indian law is vast, layered, and deeply citation-dependent. A litigator arguing before a Bench, an in-house counsel reviewing a contract, a researcher tracing a constitutional principle — each needs the same thing: the actual authority, not a paraphrase of it.

General-purpose AI tools were not built with this in mind. They hallucinate citations, conflate jurisdictions, and cannot distinguish overruled precedent from good law. That gap is not a minor inconvenience — in legal work it is a professional liability.

Niyam is our answer. An AI assistant that treats citation accuracy and good-law status as first principles, not afterthoughts. One that grounds every response in the corpus of Indian judgments and statutes rather than in statistical patterns alone.

What grounds our answers

Niyam indexes more than 72,000 Indian judgments from the Supreme Court and High Courts, cross-referenced against the statutes they interpret. When you ask a research question, the answer traces back to specific paragraphs, bench compositions, and citation chains — not to a model's best recollection.

Every answer Niyam generates identifies the source judgments it relied on. Good-law checking surfaces whether a cited case has been approved, distinguished, overruled, or criticised in subsequent decisions. You see the authority; you decide the weight.

We do not invent precedents. Where the corpus does not contain a directly on-point authority, Niyam says so. We would rather give you a narrower, accurate answer than a confident, fabricated one.

How Niyam works under the hood

Most AI tools generate text from statistical patterns learned during training, then hope the citations they produce are real. Niyam works the other way round. When you ask a question, it first retrieves the actual judgments and statutory provisions relevant to your query from its indexed corpus, then reasons over those retrieved sources to compose an answer. The authority comes first; the answer is built on it.

That retrieval-grounded design is why the citations Niyam gives you point to judgments that exist. Each proposition in an answer is tied to the paragraph, section, or rule it came from, so you can open the source and read it yourself before you rely on it. The model is never asked to remember a case from training — it is shown the case and asked to read it.

On top of retrieval sits good-law checking. Before you build an argument on a precedent, Niyam surfaces how later courts have treated it — followed, distinguished, referred to, or overruled — so the first time you learn a case has moved is not at the lectern or across the table. None of this removes your judgment from the loop. It removes the grunt work that stands between you and the moment your judgment is needed.

Built for how Indian lawyers actually work

Niyam is designed around the rhythms of Indian legal practice: bench memoranda, written submissions, opinion drafting, contract review, statutory interpretation, and matter research. The interface does not ask you to adapt your workflow to the tool.

Research outputs are structured for direct use — formatted, cited, and good-law-checked. Drafting assistance respects the conventions of Indian legal drafting. Contract review flags clauses against the relevant Act, not against generic global benchmarks.

The product is available on a credits model. You start with a ₹100 trial that grants 200 credits — enough to explore every major capability before committing further.

Who Niyam is for

Niyam is built for the range of people whose work runs on Indian statutes and judgments. Litigators who need binding authority before a hearing and persuasive authority in reserve. Law firms that want research and drafting that is fast, source-grounded, and consistent across a team. In-house counsel answering the business quickly without trading away rigour.

It is equally for those without a firm's database budget: solo practitioners carrying the whole research load alone, and law students learning to follow a proposition back to the authority it rests on. And it is for full-time legal researchers who need retrieval that surfaces the case on point even when it is phrased differently from the query.

What unites them is a refusal to rely on an answer they cannot trace. Niyam is designed for that instinct — to make verification fast, not optional, so the people who use it move quicker without lowering their standard for what counts as a reliable answer.

Our commitments

Your work product is private. Matters you research, documents you draft, and queries you run are not sold to third parties and are not used to train public AI models. Confidentiality is not a marketing claim for us — it is the foundation of legal professional trust.

We are honest about what AI can and cannot do. Niyam assists your professional judgment; it does not replace it. We design for human-in-the-loop workflows. Every answer is a starting point for your analysis, not a conclusion in itself.

We are committed to accuracy over impressiveness. When the model is uncertain, it signals uncertainty. We invest continuously in citation verification, good-law checking, and hallucination reduction — because errors in legal work have real consequences for real people.

India-first, always

Niyam is built in India, for Indian legal practice. The corpus is Indian. The statutory references are to Indian legislation — the Indian Contract Act, 1872; the Code of Civil Procedure, 1908; the Companies Act, 2013; the Constitution of India, and the full body of central and state legislation.

We are not a global product localised for India. We are an Indian product, designed with the specific demands of Indian courts, Indian statutes, and Indian legal professional culture from the ground up.

A note on professional responsibility

Niyam is a research and drafting tool. It assists advocates, counsel, and legal professionals who retain full professional responsibility for their work product. Nothing Niyam produces constitutes legal advice to end clients. The advocate using the tool is the professional; Niyam is the instrument.

We build with that boundary explicit in the product design. AI assistance that obscures professional accountability is dangerous. Ours is built to make the professional more effective, not to substitute for their judgment.