You can run a working litigation practice on the eCourtsIndia MCP with just Claude Desktop, the MCP connection, and one shared sheet. It covers roughly 70 percent of what paid case-management platforms do, tracking cause lists, new orders, and stalled matters, and the whole setup takes a single afternoon rather than a month of onboarding.
TL;DR: Pair Claude Desktop with the eCourtsIndia MCP and one Google Sheet of CNRs. Ask five plain-English questions a week, what is listed tomorrow, what orders landed, what has stalled, what is opposing counsel doing, what has the bench ruled, and you have a solo-practice tracking system in an afternoon. It will not bill, file, or give advice; for that you still buy a paid platform.
Provakil reportedly tracks 80,000 to 90,000 cause lists a day for enterprise clients. Zelican syncs with Supreme Court, High Courts, district courts, and tribunals. Komlit from Lexplosion, MyKase, ProInd LPMS all do credible work. They are the well-funded versions of the litigation-management stack. By some estimates they cost between tens of thousands and lakhs of rupees per year depending on seats and features. And they all take a month of onboarding.
A solo advocate with 140 pending matters, a Delhi High Court roll number, and a weekly WhatsApp dump to the client list can get 70 percent of what those platforms offer in one afternoon. The stack is Claude Desktop, the eCourtsIndia MCP, and one shared Notion or Google Doc. The MCP sits on top of the public eCourts record set, spanning 28.3 crore (283 million+) case records and growing across 25 High Courts and the district judiciary, so the same connection that watches one matter can watch a hundred. This post shows exactly how.
The five questions that run a litigation practice
Every matter-managing advocate asks the same five questions every week. The MCP answers each with one prompt. (Your CNR, Case Number Record, the 16-character identifier the court assigns each case, is the key the MCP uses to track every matter.)
| Question | MCP approach |
|---|---|
| What is listed tomorrow across my matters | search_causelist with your advocate name, sorted by courtroom |
| What orders came in this week | bulk_refresh_cases with your CNR list, filter orders in last 7 days |
| Which matters have not moved in 60 days | bulk_refresh_cases, compare lastHearingDate to today |
| What is opposing counsel doing on her other matters | search_cases with advocates="Opposing Counsel" |
| What has the bench ruled on similar matters recently | search_cases with the judge’s name, hasOrders=true, last 6 months |
The filter names in that table are a shorthand for how you set the workflow up in plain English. You are not memorising parameters. You describe the slice you want, an advocate, a date window, only matters that have a fresh order, and Claude assembles the call. Think of hasOrders or lastHearingDate as the questions you ask, not switches you flip. For that first question there is now a second route on the partner API: rather than searching by your advocate name, you can ask the cause-list check whether specific cases are listed. Pass up to 100 CNRs straight from your docket sheet in a single call to /api/partner/causelist/cnr/batch (a single case is just a one-element array), and it tells you whether and when each is set down and what it is listed for, without scanning a full list. That check is billed per CNR, ₹0.30 pay-as-you-go or ₹0.10 on a subscription. The MCP now wraps that same check as a native tool, check_cnr_causelist, so Claude can take up to 100 CNRs from your docket and tell you which are listed for an upcoming hearing, and where and what for, in a single step.
The one-time setup
Open Claude Desktop. Install the MCP in 90 seconds using our MCP 101 guide. For a click-by-click walkthrough, see our step-by-step guide to connecting the eCourtsIndia MCP to Claude. Paste your CNR list into a Google Sheet with columns for Matter, Client, Counsel Assigned, Exposure, Last Update. That sheet is your docket. That is the whole stack.
A full week on the MCP
A practice does not run on a single Monday prompt. It runs on a rhythm. Here is a routine that fits around hearings rather than competing with them: a heavy pull at the start of the week, light order checks in the middle, and a portfolio review on Friday before you sign off.
Monday: the cause-list pull
Monday is for knowing the week ahead. You want every listing where your name appears, grouped so you can plan travel between courtrooms and hand juniors their briefs. The setup is one instruction.
Pull the cause list for this week across all my matters. Search by my advocate name in Delhi High Court and the district courts on my CNR list. Group the result by date, then by courtroom. For each listing give me the case name, the stage, and the CNR. Put tomorrow's listings at the top.
Mid-week: the order check
Wednesday and Thursday are for catching what moved. Benches upload orders on their own schedule, so a matter heard Tuesday may only surface its order on Thursday morning. A light, targeted refresh keeps you from being the last to know. When you need the full paper trail on a single matter rather than a portfolio sweep, list_case_orders returns every order and judgment on that case, newest first, and pairs naturally with get_order_markdown when you want the text of a particular one.
Refresh only the matters that were listed in the last three days. For each one, tell me whether a new order has been uploaded since Monday. If yes, give me a two-line plain-English summary and the next listing date. If a matter was listed but shows no order yet, flag it as "awaiting order".
Friday: the portfolio review
Friday is for the matters nobody is talking about. The cases that quietly stall are the ones that turn into an angry client call three months later. The portfolio review surfaces them before they fester.
Review my whole docket from the Google Sheet. Refresh all CNRs. Flag every matter with no movement for 60 or more days. For each flagged matter, give me the last recorded event and how many days ago it was. Sort the flagged list by exposure, highest first, using the Exposure column in my sheet. End with a one-line count of healthy vs stalled matters.
You can also have the Monday pull run on a schedule rather than by hand. The workflow is set up so the docket sheet is read, every CNR refreshed, and the summary posted to your team space at a fixed time, the way the original weekly routine below was written.
Every Monday at 07:00 IST: 1. Read the CNR list from my Google Sheet. 2. Refresh all of them via bulk_refresh_cases. 3. For each CNR, report: last order date, last hearing date, next listing, current status. 4. Flag any matter where no movement for 60+ days. 5. Flag any matter with a new order in the last 7 days. 6. Summarise in a table. 7. Paste the table into my shared team Notion at /litigation/weekly.
One practical limit to build into the routine: a single bulk_refresh_cases call handles up to 50 CNRs, so a 140-matter docket is refreshed in three batches rather than one, and Claude chunks the sheet for you. If you would rather watch the whole portfolio as one unit than refresh it list by list, the MCP also exposes a monitor_portfolio tool for programmatic portfolio monitoring that lives in your Claude window and your sheet. All of these, check_cnr_causelist, bulk_refresh_cases, list_case_orders, and monitor_portfolio, are part of the MCP’s current 22-tool set. Note the distinction: the eCourtsIndia website itself has no aggregate “My Cases” dashboard, so the consolidated view is something you assemble through the MCP, not a page you log into.
For High Court and district court matters, that per-case pull now surfaces richer detail than it used to. Alongside the order, hearing, and listing dates, recent records expose fields like the e-filing number and date, filing type, litigant type, and structured parties, and for district matters the interlocutory applications, transfers, and full hearing history. These appear on the case page itself; you do not have to ask for any of it, and when a court publishes it, the refresh captures it.
Ten minutes from Monday 7 am. The team stand-up has an agenda. The client emails have subject lines. The adjournment narrative is not a surprise.
What a consolidated cause list looks like
The point of the Monday pull is that you stop opening a dozen court websites. Instead of checking the Delhi High Court list, then three district court lists, then a tribunal, you get one table for the whole week. A worked output reads something like this.
| Date | Court / Room | Matter | Stage | CNR (last 4) |
|---|---|---|---|---|
| Tue | DHC, Court 12 | Mehta v. Union of India | Final arguments | …4471 |
| Tue | DHC, Court 12 | Verma Estates v. DDA | Reply filed | …8820 |
| Wed | Saket, Court 3 | State v. Khanna | Charge | …1093 |
| Thu | Patiala House, Court 7 | Rao v. Sharma | Evidence | …5512 |
| Fri | NCDRC | Anand v. BuildWell | Admission | …3076 |
Two matters in the same room on the same morning means one trip, not two. A charge hearing and an evidence hearing on consecutive days means you brief the right junior for each. The consolidation is the whole value: the information always existed on the public record, but no human was going to assemble it by hand every Monday. If you want to see the raw lists behind a pull, the eCourtsIndia cause list page carries the same daily data.
Integrations you will want in week two
Once the pulls are reliable, the next gain is getting the output to where your team and clients already are.
- Push the weekly summary to a shared Slack or WhatsApp channel so juniors and partners see the same docket you do.
- Create a Google Calendar event for every “listed tomorrow” row, with the case name and courtroom in the title and the CNR in the notes.
- Generate a short client-facing WhatsApp line per matter, “Your case is listed on Thursday at Patiala House for evidence,” and review the batch before it goes out.
- Maintain a separate CSV for audits, so you can demonstrate compliance with your firm’s client-communication policy.
The WhatsApp client update is the one that earns goodwill fastest. Clients rarely chase you because they have nothing to chase; a two-line note the evening before a listing tells them you are on it. Keep a human in the loop on those messages. The MCP drafts, you approve, the client receives. Never wire an automated send straight to a client list without a review step.
DIY MCP versus paid practice-management tools
This is not a case of one approach beating the other. They sit at different points in a practice’s life. The honest comparison is about where each one earns its keep.
| Capability | DIY MCP stack | Paid platform |
|---|---|---|
| Cause-list and order tracking | Yes, on the public record | Yes, often with denser daily coverage |
| Stalled-matter alerts | Yes, one prompt | Yes, dashboards and reminders |
| Time capture and billing | No | Yes |
| Trust accounting | No | Yes |
| Client portal | No, you push updates manually | Yes |
| Conflicts checking | No | Yes |
| Immutable audit log | Only what you build into a CSV | Yes, built in |
| Setup time | An afternoon | Reportedly around a month |
| Cost | Your existing Claude subscription | By some estimates tens of thousands to lakhs a year |
Read that table fairly in both directions. The paid platforms are not overpriced; trust accounting and conflicts checking are genuinely hard problems and worth paying for once they matter. The MCP stack is not a toy; for a solo or a two-person chamber whose pain is tracking, not billing, it does the job that actually keeps clients calm. Pick the tool that fits the problem you have today.
Security and data handling
A practice tool touches client information, so a few habits are worth setting on day one. The case data the MCP reads is already public record on the eCourts system, but your docket sheet, your client phone numbers, and your internal notes are not, and they deserve care.
- Keep the docket sheet on an account you control, with sharing restricted to the people who need it. The CNR list itself is low-risk; the client and exposure columns are not.
- Store CNRs and case numbers in the sheet. There is no reason to keep client contact details in the same place you run automated pulls if you can avoid it.
- Treat any client-facing message as a draft until a human approves it. This is a security control as much as a courtesy, since it catches a wrong number or a mismatched matter before it leaves.
- If your firm has a client-communication policy, keep the audit CSV so you can show what was sent, to whom, and when.
- Review who has edit access to the shared sheet every few months. Stale access is the most common quiet leak in a small chamber.
What the MCP will not do
An honest tool comes with an honest list of limits. The MCP reads and organises the public court record. It does not replace your judgment, and there are things it simply cannot do.
- It does not file anything. No e-filing, no vakalatnama, no replies. It watches the record; it does not write to it.
- It does not capture your billable time or raise an invoice. If billing is a revenue leak, that is a paid-tool job.
- It is only as current as the court’s own uploads. If a bench has not uploaded an order, the MCP cannot show you an order that is not there yet.
- It does not give legal advice or decide strategy. A plain-English summary of an order is a starting point for your reading, not a substitute for it.
- It does not guarantee completeness on records the public system itself does not expose. Always confirm anything consequential against the source.
None of that undercuts the value. Knowing the limits is what lets you trust the parts that work.
How this compares to paid products
Paid products have real advantages. Time capture. Billing. Trust accounting. Client portals. Conflicts checking. Document assembly. Firm-wide reporting. You will outgrow this DIY stack when your practice becomes a team of five or more, when billable-hour capture becomes a revenue issue, or when a client audit demands an immutable log. That is fine. The MCP stack gets you from zero to a working practice tool. Pay for the rest when you need it.
What this means for advocates
You do not need a year-long procurement cycle to get your practice organised. Start with one matter. Paste the CNR into Claude and ask “What happened in this case in the last year. Explain in English.” If you like the answer, expand to ten. Then to a hundred. Our CNR Decoded post explains how the 16-character identifier you are about to paste actually works.
A working practice tool used to cost a month of onboarding and a five-figure budget. The version that covers 70 percent of the work now costs an afternoon.
Further reading: The 5 AM Cause List Problem, Case Status Dictionary.
Frequently Asked Questions
What do I need to run my litigation practice on the eCourtsIndia MCP?
You need three things: Claude Desktop, the eCourtsIndia MCP connection, and one shared Google Sheet or Notion page holding your CNR list. That stack covers roughly 70 percent of what paid platforms do. Start your matter searches at eCourtsIndia and paste the CNRs into your docket sheet.
How do I check what is listed tomorrow across all my matters?
Use the search_causelist function with your advocate name and sort by courtroom. The MCP pulls every listing where you appear, so you are not checking cause lists court by court. You can also browse daily lists directly on the eCourtsIndia cause list page.
How does the MCP stack compare to paid tools like Provakil or Zelican?
Paid products reportedly track 80,000 to 90,000 cause lists a day and add billing, trust accounting, and client portals, but by some estimates cost tens of thousands to lakhs per year and take a month to onboard. The MCP stack gets you working in an afternoon. See our MCP 101 guide.
How do I spot matters that have stalled?
Run bulk_refresh_cases across your CNR list and compare each matter’s last hearing date to today, flagging anything with no movement for 60 or more days. The MCP does this in one prompt every Monday. Look up a litigant’s cases anytime on the public eCourtsIndia litigant pages.
When should I move to a paid practice-management platform?
Move on when your practice becomes a team of five or more, when billable-hour capture turns into a revenue issue, or when a client audit demands an immutable log. Until then the MCP stack covers the core work. Research opposing counsel and bench patterns at eCourtsIndia.
Is my client data safe when I run these pulls?
The case data the MCP reads is already public record on the eCourts system, covering 28.3 crore (283 million+) records and growing across 25 High Courts and the district judiciary. Your docket sheet, client numbers, and notes are private, so keep the sheet on an account you control, restrict sharing, and treat every client-facing message as a draft until a human approves it.
What will the MCP not do for my practice?
It does not file pleadings, capture billable time, raise invoices, or give legal advice, and it is only as current as the court’s own uploads. It reads and organises the public record so you can act on it; it does not replace your judgment or write to the court system.