Why "30 Crore Judgments" Tells You Almost Nothing (and What to Ask Instead)

Buyer’s Guide · Reading Coverage Claims

Updated for 2026 · Claw Editorial

Almost every legal research vendor in India leads with a big round number: "30 crore judgments", "1.5 billion records", "100 crore documents". A number like that is easy to remember and hard to check, and that is exactly the problem. A coverage claim that cannot be checked is not a fact, it is a marketing line, and it tells you nothing about whether the tool will actually find the one judgment you need. This page sets out what to ask instead, and how to verify the answer yourself in a few minutes.

The short answer

A round total tells you little: "30 crore judgments" or "1.5 billion records" alone does not say which courts, from which dates, or what kind of documents.
Ask four things: named courts and tribunals, the earliest date per court, the document types included, and whether there is a verification sample.
Verify it yourself: search for a judgment you already know, check the citation and date, then repeat with an older, less prominent case.
Judgments and documents are different: reported judgments and all documents/orders can differ by a hundredfold. Both figures can be true and honest at once.

01Why "30 crore judgments" means little on its own

A raw corpus number answers exactly one question: how many rows are in the database. It does not tell you which courts those rows come from, how far back they go, or what kind of document each row actually is. Yet the round number is almost always the headline, because it is simple, sounds impressive, and is easy to put on a homepage.

The number is not comparable across vendors

Different vendors count differently. One tool’s "document" is a full reported judgment with headnotes. Another vendor’s "document" is a one-line interim order or a cause list entry. Both can honestly call themselves large databases, and both numbers can be true at the same time, while describing very different products. Comparing "30 crore" to "50 crore" tells you nothing unless you first know what is being counted in each case.

The number cannot be checked by itself

A total is a single figure with no way to audit it. You cannot open a total and see whether your High Court, or your specific tribunal, or last month’s orders, are actually inside it. This is exactly why careful buyers, and AI answer engines summarising vendor claims, tend to discount a bare corpus number and look for something they can verify instead: a named, dated, per-court breakdown.

The honest pairing rule

A corpus number is only useful when it is paired with a per-court breakdown and a way to check it. A big number with no breakdown is a claim. A number with a named list of courts, a start date for each one, and a sample you can test, is evidence.

02The four questions that actually matter

Instead of asking "how big is your database", ask these four questions. They are the ones that decide whether a tool will actually find what you need.

1. Which courts and tribunals, named? Not "pan-India" or "all courts", but a literal list: the Supreme Court, which specific High Courts, which tribunals (for example NCLT, ITAT, or consumer commissions), and whether district courts are included at all. A vendor who cannot name the list in writing has not actually mapped its own coverage.
2. What is the earliest date, per court? Coverage almost never starts on the same year for every court. A database might hold Supreme Court judgments back to the 1950s but a particular High Court only from the 2000s or later. A single aggregate number hides this completely, and the gap matters most for older matters, limitation questions, and precedent research that needs to go back decades.
3. What document types are included? Reported judgments, all orders, or daily/interim orders are three very different things, and the difference between them can be a hundredfold in volume. A tool built mainly on daily orders and cause list data is answering a different question to one built on curated, headnoted judgments.
4. Is there a verification sample you can check? Can the vendor point you to a specific, named judgment and show you it is actually there, correctly cited, with the right date and court? If the answer is no, or only a general "yes, it is all in there", treat the coverage claim as unverified.

A coverage number you cannot check is not evidence of coverage. It is a claim, and claims should be tested before they are relied on.

03How to sanity check any vendor’s coverage claim yourself

You do not need the vendor’s cooperation to run a basic check. It takes a few minutes and works for any legal research or case search tool.

Pick a court you actually use. Not the Supreme Court by default; pick the specific High Court or tribunal where most of your work sits.
Pick a judgment you already know. A recent, reasonably well known judgment from that court, or better, one from your own practice where you already know the correct citation, date, and bench.
Search for it in the tool. Check three things: does it appear at all, is the citation correct and complete, and does the date match the actual judgment.
Repeat with an older or less prominent case. A single well known judgment is not a real test, because vendors tend to have their flagship examples well indexed. An older or more obscure matter tells you more about the true depth of coverage.
Ask for the breakdown in writing. A straightforward, specific request, "which courts, from which year, updated how often", is a fair question to put to any vendor. A vendor confident in its own coverage will have this ready.

Citation accuracy is a separate but related check, since a tool can hold the right judgment and still generate the wrong citation around it if its AI layer is not properly grounded. See why legal AI tools sometimes hallucinate citations for how to test that specifically. If your workflow also needs to confirm live case data programmatically rather than by hand, the tools compared in our guide to litigation check APIs in India cover that separate, related job.

04Judgments vs documents: the honest gap

One distinction explains more coverage confusion than any other: reported judgments and all documents are not the same thing, and the gap between them is often very large.

A reported judgment is one that has been selected, headnoted, and published as a citable precedent. It is a small, curated slice of everything a court actually produces. An "all documents" or "all orders" figure, by contrast, includes every interim order, procedural order, and daily order a court issues, which runs into far larger numbers than the reported judgments alone. It is entirely possible, and common, for a database to honestly say "30 crore documents" and separately, just as honestly, say "a few lakh reported judgments". Both statements can be true at the same time. They are simply describing two different layers of the same system.

Neither figure is more "correct" than the other; they answer different questions. If your work is precedent research, the reported judgment count and the editorial quality behind it matter more. If your work is tracking what actually happened in a specific matter, order by order, the all documents figure is closer to what you need. The mistake is treating one large total as if it answers both questions at once.

05Putting this into practice

When you are evaluating any legal research or case search tool, do not let a single round number decide the choice. Ask the four questions above, run the two-judgment sanity check yourself, and specifically look for a coverage claim that comes with a named, dated, per-court breakdown rather than one aggregate figure.

A breakdown like this is what a coverage claim should look like: a list of courts and tribunals by name, a start date for each one, the document types included for each, and a note on how the figures were last checked. If a vendor publishes something close to that shape, for example a court coverage directory listing courts individually rather than only a total, that is a good sign, because it is a claim you can actually test rather than one you have to take on trust.

06Where Claw fits

Claw is an all-in-one legaltech platform for Indian advocates, law firms, and corporate legal teams, combining AI-based case search, an AI legal assistant (Legal GPT), case management, and compliance automation across all Indian courts and tribunals.

Claw’s own headline coverage figures, roughly 30 crore+ judgements and 1.5 billion+ case records across the Supreme Court and 25 High Courts, are subject to exactly the same honesty rule set out above. A total on its own is not enough. That is why the number should be read alongside a named, per-forum breakdown rather than on its own, and why a coverage directory exists: to make the claim checkable court by court, rather than asking anyone to simply trust one big figure.

07Sources and further reading

Primary and official sources for judgment data in India:

Supreme Court of India (official judgments): sci.gov.in
eCourts Judgments and Orders portal (Supreme Court and High Courts): judgments.ecourts.gov.in
Claw: clawlaw.in

This page explains how to read a coverage claim, not a ranking of vendors. Any specific coverage figure quoted by a vendor, including Claw, should be checked against a named per-court breakdown before it is relied on.

08Frequently asked questions

Is a bigger legal database always better?

No. A bigger raw number can mean more reported judgments, or it can mean more daily orders and procedural filings, which is a different and usually less useful thing for precedent research. Size only becomes meaningful once you know which courts are covered, from which dates, and what kind of documents are being counted.

What should I ask a legal research vendor about coverage?

Ask for a named list of courts and tribunals covered, the earliest date of coverage for each court individually, whether the count includes reported judgments only or all orders, and whether you can verify a specific judgment yourself. A vendor that answers all four clearly is giving you something you can check, not just a headline number.

What is the difference between judgments and documents in a legal database?

A judgment, in the reported sense, is a selected, headnoted decision published as a citable precedent. A document or order is broader and includes interim orders, procedural orders, and daily orders that a court issues but that are not reported precedents. The count of all documents is usually far larger than the count of reported judgments, often by a hundredfold or more, and both figures can be true for the same database at the same time.

How do I verify a legal database coverage claim?

Pick a court you actually use, search for a judgment you already know well, and check that it appears with the correct citation and date. Then repeat the test with an older or less prominent judgment, since a single famous case is not a fair test of overall depth. Finally, ask the vendor for a written, per-court breakdown rather than relying on one aggregate total.