The Economist recently advised its readers to keep a glass of water and 1,000 mg of paracetamol within reach before visiting an Indian government website. The piece caused a fair amount of outrage in India. The publication must thank its stars that it is unlikely to get sued because then its owners would have had to navigate Indian court websites. The visa portal that The Economist was mocking is, by comparison, a spa.
Over five crore cases lie pending in India’s courts — the most written-about number in Indian law. There are also shelves of committee reports on ‘access to justice’. Legal scholars have commented on fairness, bias, poor judgments, and now the use of artificial intelligence (AI). This article will focus on access to data. Not the kind of data researchers need for policymaking or optimising court operations — that’s a different battle, for another day. The data in question here is more banal, but far more important. It is the basic information that a litigant or a lawyer needs to follow their case in a court.
The journey of a court website user
Let’s assume you are a litigant — a small company that has sued a large buyer for non-payment of bills, an employee who has sued their employer for unpaid salary or a senior citizen who is fighting the government in court for pension. Let’s assume you’ve been told that your case is likely to be heard tomorrow. What do you do the previous day?
Perhaps you want to find out where it stands. Is it listed? Who is the judge? Will it actually be taken up? What happened at the last hearing? You can call your lawyer to check, and chances are the average lawyer is also struggling with the same questions. The scramble begins around 6 pm, because that is when the next day’s cause list — the court’s daily schedule — is finalised and uploaded. Or should be. You enter the web address and wait. The page loads endlessly, or times out, or produces a smudged captcha that must be copied out carefully before the page dies anyway.
Now you face an almost metaphysical question posed by the Indian court website: is the server down? Did I get the captcha wrong? Should I refresh? Or wait? You do what everyone does — refresh, retype, refresh — at the precise hour when every other litigant or lawyer in the state is doing the same thing.
The average court website experience
Over the last few months, we ran a small experiment, pointing an automated monitor at three websites and checking them every 15 minutes. The National Company Law Tribunal (NCLT)’s portal was unreachable on roughly one in five visits. There’s a running joke among practitioners that the website seems almost human in its behaviour—it works well on a Monday morning, but by midweek it is burned out and exhausted. The Bombay High Court’s website, serving the country’s busiest high court, failed nearly four times as often during court hours as it did overnight.
Once it loads, finding the case details isn’t easy. Users must possess specific information such as the registration number or the filing number. To search by litigant or lawyer, one often needs to specify the exact court complex the case is in. In a system where neither registration numbers nor spellings are standarsized, a basic search feature that we generally take for granted in a modern website is the first barrier to access.
We recently surveyed a handful of lawyers, researchers and litigants about their experience with Indian court websites. It’s a small sample of around 70 respondents, but they raised similar concerns. On the basic tasks involved in using a court website, the failures were routine: for a majority of respondents, captchas, over-demanding search forms, erroneous responses and session time-outs were problems at least half the time. Nearly 60 per cent said such glitches disrupted their workflow, while more than half said they impeded access to critical court information. One of our more disgruntled respondents captured the indignity and tedium of it all rather eloquently: “During important work, I really don’t want to waste time identifying which images have traffic signals or bicycles or bridges in them before I can log in. Surely better alternatives exist.”
Multiple sources of truth, all incomplete, incorrect or outdated
Most courts in India have two independent websites. There is a national portal for district courts, and each district court also has its own website, built by the same National Informatics Centre, yet with slightly different navigation. There is an integrated high court services portal, and the Bombay, Delhi and Madras high courts each run their own websites, which do not always agree with it. The cause list may live on one site, case status on another, and orders on a third. These sites are redesigned every few years, and old links disappear without notice.
Suppose our intrepid explorer—lawyer or litigant—gets through this maze. The data that greets them is sometimes stale, and some of it is outright wrong. Yesterday’s hearing may surface next week; cases disposed of months ago may still appear as “pending”. Orders from someone else’s case may be uploaded against yours. For years, the older version of the NCLT website cluttered case pages with “connected matters” that had no connection at all.
This has been measured and written about. When researchers Devendra Damle and Tushar Anand systematically analysed the e-Courts data in 2020, they found missing fields, malformed entries and outright incorrect data wherever they looked.
Our experience has been similar. In one instance, we observed the court uploading an order relating to a civil writ petition against a criminal writ petition with the same number. In another, the website simply added judges to the roster whose names do not appear on the orders. Hallucinations in Indian court data are not the fault of LLMs alone.
Easy fixes, uneasy incentives
There is an infamous story from two decades ago about Google testing 41 shades of blue before settling on the right one. It sounds absurd, but there is a logic to it: Google operates in a ferociously competitive market where even a marginal improvement at the fringes is worth real money. A court website is the opposite. It is a monopoly with a perfectly captive audience. Litigants and lawyers can yell themselves hoarse, but there are few people listening. And even if some kind-hearted judge or justice were to take notice, they would themselves be at the mercy of the NIC.
This may be the good part of the system. At least there are attempts to standardise reporting across district courts, high courts and the Supreme Court through the National Judicial Data Grid. Beyond that lies the Wild West of tribunals and consumer courts that sit outside this infrastructure. Tracing a dispute across multiple courts is difficult because the moment it is transferred—even within the same court complex—the case is renumbered. Follow a case from its first hearing, through trial, through the high court appeal and on to the Supreme Court, and you begin to understand why Camus imagined Sisyphus happy.
This is the state of court websites in 2026. In a country that builds software for some of the world’s finest companies, a co-equal branch of government runs on websites that would make even experienced bureaucrats scream. So our lawyer, nearing midnight, does what lawyers and litigants across the country do: she stops trusting what is on the screen and calls the court clerk.
Siddarth Raman is co-founder and CTO at The Professeer. Views are personal.
(Edited by Prashant Dixit)