Search for the best legal research tool in India and you will get a page of listicles that all say the same thing: nine platforms, each described in two flattering sentences, no sources, no dates, and no indication that the author opened any of them. The numbers in those posts are usually lifted straight off the vendor's own homepage and then repeated for years after the vendor quietly changed them.
This is an attempt at the opposite. Every factual claim below carries the URL it came from and the date it was checked. Where a vendor does not publish a number, this post says that the number is not published, rather than filling the gap with an estimate. Where a competitor is genuinely better than CourtMesh at something, it says so, by name, in its own section.
Disclosure, up front
CourtMesh publishes this post and CourtMesh is one of the tools listed. You should discount our own entry accordingly. The reason every other entry carries a source URL is so that you do not have to take our characterisation of a competitor on trust: open the link and check. If you find something here that is wrong or out of date, tell us at contact and we will correct it.
How to Read This List
Sources were last checked on 18 September 2026. Legal technology moves quickly and several of these vendors changed materially in the first nine months of 2026, so treat any claim here as a snapshot with a date on it rather than a standing fact.
Three distinctions matter more than anything else on a feature grid, and almost every list in this category blurs all three.
Judgments are not dockets
A reported judgments publisher holds decided opinions, usually appellate, usually edited. A docket source holds the case as it moves: filings, case status, order sheets, listing dates, parties. These are different corpora that answer different questions, and a platform strong at one is often absent at the other.
Indexed is not analysed
A platform can hold a hundred million records and have run a model over a small fraction of them. Ask for both numbers separately. Any vendor quoting one number for both is either not measuring or not telling.
Editorial is a real product
Headnotes, catchwords, a curated citator and a firm view on whether a case is still good law are human editorial work built over decades. No amount of retrieval quality substitutes for it, and the platforms that have it are genuinely ahead on it.
Indian Kanoon
The default starting point for most Indian legal research, and the one entry on this list that is free to use without an account. Its browse page lists the Supreme Court, 24 High Courts and more than 20 tribunals, and states a corpus of over 3 crore orders (https://indiankanoon.org/browse/). District Court coverage is the notable limit: two district courts, Delhi and Bangalore.
A paid membership exists for consumers at 500 rupees a month or 5,000 rupees a year, with a one month free trial (https://indiankanoon.org/members/faqs/). There is also a published API rate card, which is unusual in this market: 0.50 rupees per search request, 0.20 rupees per document, 0.05 rupees per document fragment and 0.02 rupees per document metainfo call, with 500 rupees of free signup credit (https://api.indiankanoon.org/pricing/). API use carries a mandatory attribution condition: a Powered by IKanoon badge must be displayed, and the terms extend that requirement to retrieval augmented generation and fine tuning use (https://api.indiankanoon.org/terms/).
Where Indian Kanoon is genuinely stronger than CourtMesh
It is free, it needs no account, it has been the profession's shared reference for well over a decade, and citing an Indian Kanoon link to another lawyer requires no explanation of what the site is. That trust is not a feature anyone can ship; it was earned. If your research question is answered by a reported Supreme Court or High Court judgment, starting anywhere else is usually wasted effort.
One live development is worth knowing before you build a workflow on it. On 29 May 2026 the Delhi High Court, ruling across a set of right to be forgotten petitions, directed Indian Kanoon to restrict name based search for specified records while leaving those judgments reachable by case number, citation, court and date. Indian Kanoon has appealed and the order stands pending appeal; at a hearing on 17 September 2026 the court observed that disabling name search means, for practical purposes, the judgment has been blocked (https://www.medianama.com/2026/06/223-google-de-indexing-indian-kanoon-search-restrictions-right-to-be-forgotten/ and https://www.medianama.com/2026/09/223-delhi-hc-indian-kanoon-name-searches-right-to-livelihood/). This affects every platform that offers party name search in India, CourtMesh included.
SCC Online
Part of the EBC Group, the publisher behind the Supreme Court Cases reports (https://www.scconline.com/about-us). SCC Online is the reference most Indian litigators cite from, and its value sits in editorial work rather than raw volume: headnotes, catchwords, curated citation treatment and the SCC citation itself.
On 15 January 2026, Harvey and SCC Online announced a partnership making SCC Online content selectable as a knowledge source inside Harvey. The announced scope is case law, legislation, secondary materials, journals, news, lectures, transcripts, foreign content and legal forms (https://www.harvey.ai/blog/harvey-partners-with-scc-online). Neither release states exclusivity. As at 18 September 2026 SCC Online does not publish a corpus document count anywhere we could find, so this post does not state one.
Where SCC Online is genuinely stronger than CourtMesh
Editorial depth and citator authority, and it is not close. A headnote written by an editor who read the judgment, a considered view on whether a case has been overruled, distinguished or followed, and a citation a judge recognises on sight are things CourtMesh does not have and does not claim. If your work is appellate advocacy where the authority of the citation matters as much as finding the case, SCC Online is doing something we are not.
Manupatra
The other established Indian research publisher, with a comparable editorial and citation proposition. Manupatra's own legal research page claims more than 10 million legal documents and judgments, and a citation database drawing on more than 320 Indian journals (https://www.manupatra.ai/legal-research).
On the same day as the Harvey and SCC Online announcement, 15 January 2026, Legora and Manupatra announced a partnership that Legora describes as exclusive, covering case law, regulations, commentary and citations (https://legora.com/newsroom/legora-and-manupatra-announce-exclusive-partnership and https://www.artificiallawyer.com/2026/01/20/legora-partners-with-indian-legal-data-company-manupatra/). Two of the largest international legal AI platforms bought Indian judgment content from the two largest Indian judgment publishers in the same week, which tells you something about where the scarcity in this market actually is.
Where Manupatra is genuinely stronger than CourtMesh
The same answer as SCC Online: editorial headnotes and a citator built over decades, plus journal and commentary coverage across more than 320 titles that CourtMesh has no equivalent of. Secondary literature is a real research input and we do not carry it.
CaseMine
Built by Gauge Data Solutions, CaseMine covers Indian, US and UK case law and ships an AI assistant branded AMICUS, with its cross jurisdictional expansion reported by Bar and Bench (https://www.barandbench.com/law-firms/from-india-to-global-legal-research-how-casemine-is-expanding-cross-jurisdictional-access). Its visual citation mapping is the feature it is best known for among Indian researchers.
We could not verify a corpus size. The CaseMine site returned an HTTP 403 to our fetch on 18 September 2026, so no document count, court list or price appears in this entry. That is a gap in our checking, not a criticism of the product.
vLex and Vincent AI
vLex was acquired by Clio in a transaction Clio announced as completed on 10 November 2025 (https://www.prnewswire.com/news-releases/clio-completes-landmark-1b-vlex-acquisition-and-announces-500m-series-g-funding-round-at-5b-valuation-302610111.html). Its Indian site does carry Indian case law: checked directly on 18 September 2026, https://vlex.in/ lists the Supreme Court, High Courts, the State and National Consumer Disputes Redressal Commissions, Debt Recovery Tribunals, Appellate Tribunals, the Central Administrative Tribunal, the Armed Forces Tribunal, the Central Information Commission, the Authority for Advance Rulings, the Central Electricity Regulatory Commission, the Company Law Board, the Competition Commission of India, the Securities and Exchange Board and the Trademark Tribunal.
That page states no document count and no court count, so this post states neither. Note the shape of what is listed: appellate and tribunal decisions, without a docket, case status or cause list layer visible on the page. Vincent AI is the research assistant vLex sells on top of it.
LexisNexis India
LexisNexis launched Protege Workflow in India on 11 April 2026, alongside Case Analysis India, which it had introduced in February 2026 on Lexis Advance India (https://www.lexisnexis.com/blogs/in-legal/b/law/posts/launch-protege-india). Protege Workflow is described as task automation, draft generation, summarisation and custom workflows inside a private environment.
Be careful with what is often written about this launch: it is a workflow product plus a case law research tool, not a launch of Lexis+ AI in India. We could not verify what Indian primary content LexisNexis holds, which law report series it owns, or what it charges in India, so none of that appears here. The LexisNexis India site did not return a readable inventory to us on 18 September 2026.
The Indian AI Native Platforms
A cluster of Indian startups now sell AI first legal research. The pattern across all of them is that almost none disclose where their Indian case law comes from, which is itself the most useful thing to know about the category.
Jhana.ai describes a national legal archive of more than 16 million documents and has raised 15 crore rupees, with investors including Razorpay, Jio and Together Fund (https://www.jhana.ai/). Lexlegis.ai sells an assistant branded MIRA and grounds it in what it calls proprietary, verifiable legal sources; it publishes no document count and no source breakdown on its own site, so this post quotes no figure for it (https://lexlegis.ai/). Jurisphere states more than 15 million documents processed and raised 2.2 million US dollars in May 2026 (https://jurisphere.ai/). Bharat.Law claims more than 15,000 courts and more than 10 million documents (https://bharat.law/).
LegitQuest is the Indian platform whose published claims sit closest to CourtMesh's own territory: its site claims more than 500 million records across more than 10,000 courts with more than 500,000 new records daily, alongside a litigation due diligence product branded LIBIL, and it states ISO 27001 certification and CERT-In empanelment (https://www.legitquest.com/). Those are vendor claims, stated on the vendor's site, and we have no way to audit them; the same caveat applies to every volume figure on this page including ours.
CourtKutchehry sits at the smaller, freemium end with 1.18 million judgments (https://www.courtkutchehry.com/).
CourtMesh
Our own entry, held to the same standard. CourtMesh indexes 315.6 million court records sourced directly from official government portals, principally eCourts, the NJDG and court registries, with no reseller in the chain. Keyword search runs across that whole corpus. AI semantic search runs across a much smaller subset, roughly 2 million judgments, because embedding and analysing a nine figure corpus is a compute bill nobody in this market has actually paid.
The thing CourtMesh has that the judgment publishers on this list do not is the docket layer: District Court filings, live case status, order sheet timelines and party level litigation history, across the Supreme Court, all 25 High Courts, the district judiciary and tribunals. That is a different corpus from a reported judgments archive, and it answers a different question. If your question is what has been held on section 34 of the Arbitration and Conciliation Act 1996, a judgments publisher is the right tool. If your question is what is pending against this counterparty and when is it next listed, it is not.
Where CourtMesh is genuinely weaker
No editorial headnotes. No curated citator, so nothing on this platform will tell you authoritatively that a case has been overruled. No journals or commentary. No decades of brand recognition with the bench. AI analysis covers a small fraction of the corpus, not all of it. And coverage is not uniform: District Court records are thinner than High Court records, and older records are thinner than recent ones everywhere. If any of those matter more to your work than docket reach, buy something else, or buy both.
The Free Sources Worth Knowing About
Two of these are not products and will not appear on a vendor listicle, which is exactly why they belong here.
The government's own portals are authoritative and free. eCourts Services carries CNR lookup, case status, orders and cause lists (https://services.ecourts.gov.in/ecourtindia_v6/) and the judgments portal offers full text search across Supreme Court and High Court judgments (https://judgments.ecourts.gov.in/). Both put a CAPTCHA on every search and neither offers bulk download or a public API, which is the practical reason an aggregation layer exists at all. The NJDG publishes statistics rather than documents (https://njdg.ecourts.gov.in/njdg_v3/).
Development Data Lab publishes 81.2 million Indian district and sessions court cases covering 2010 to 2018, scraped from eCourts and anonymised, under a CC BY-NC-SA 4.0 licence that requires a separate licence for commercial use (https://www.devdatalab.org/judicial-data). For academic work on Indian district court litigation this is an extraordinary resource and it costs nothing. Its limits are that it stops in 2018, it is de-identified, and it is district and sessions courts only.
At a Glance
Every cell below is a claim published by the vendor on the page cited in its section above, checked on 18 September 2026, not an independent audit. Where a vendor publishes no figure, the cell says so. Nothing here is estimated.
| Platform | Stated corpus | Docket layer | Editorial headnotes | Public pricing |
|---|---|---|---|---|
| Indian Kanoon | 3 crore+ orders | 2 district courts only | No | Yes, consumer and API |
| SCC Online | Not published | Not offered | Yes | Not published |
| Manupatra | 10M+ documents | Not offered | Yes | Not published |
| CaseMine | Not verified, site returned 403 | Not verified | Not verified | Not verified |
| vLex India | Not published | Not visible on the India page | Not verified | Not published |
| LexisNexis India | Not verified | Not verified | Not verified | Not published |
| Jhana.ai | 16M+ documents | Not stated | Not stated | Not published |
| LegitQuest | 500M+ records claimed | Claimed via LIBIL | Not stated | Not published |
| Bharat.Law | 10M+ documents claimed | Not stated | Not stated | Not published |
| CourtKutchehry | 1.18M judgments | Not offered | No | Freemium |
| CourtMesh | 315.6 million records | Yes, the core of the product | No | Yes, published |
What We Could Not Verify
Listing these is the point of the exercise. A comparison that has an answer for every cell is a comparison somebody made up.
Choosing Between Them
The honest answer is that most Indian practices end up using two of these, not one, because the category contains two different products wearing the same label.
Start with what your question actually is
If you are looking for authority on a point of law, you want a judgments publisher with editorial depth. If you are looking for what is happening in a live matter or what litigation a party is carrying, you want a docket source. Buying the wrong category and then complaining about coverage is the most common failure in this market.
Run the same ten queries everywhere
Take ten real questions off your own files, including at least two District Court matters and one tribunal matter, freeze the list, and put the identical ten to every candidate before anyone shows you a price. Demos are built on Supreme Court constitutional material because that is the easiest thing in the domain.
Ask for the two corpus numbers separately
How many records are retrievable, and how many have actually been analysed by a model. A vendor who quotes one number for both has answered a different question from the one you asked.
Check the commercial terms before the technology
Published pricing, credit expiry, rollover, and what happens to unused balance on a plan change. These predict more about year two than any feature comparison predicts about week one.
Verify against the official record before you act
Every platform on this list, free or paid, ours included, is a view of what registries published. Before advising a client or relying on a case status in a filing, confirm it against the record of the court concerned.
A vendor who tells you exactly where their coverage thins out has understood their own pipeline. A vendor who tells you coverage is complete has either not looked or is hoping you will not.
Check our entry the way you would check theirs
CourtMesh indexes 315.6 million court records from official government portals, with keyword search across all of it and AI semantic search across a roughly 2 million judgment subset. What we do not have is editorial headnotes or a citator, and this post says so in our own section. Read the product detail at legal research software, the published plans at pricing, and the corpus and freshness detail at the API overview. Then run your ten queries against us and against everyone else on this page.
Explore CourtMesh
