Indian Court Arguments & Drafts for LLM Training

India’s e-courts ecosystem now makes large volumes of pleadings, written submissions, paper-books, affidavits, petitions and drafting templates openly obtainable—if you know where to look and how to extract them. Below is a detailed, step-by-step blueprint for sourcing, downloading, processing and legally using these high-value materials to fine-tune a Legal-Research LLM or populate a VectorDB.

Overview

Most Indian litigants file 6 core “argument” artefacts:

1. Petitions / plaints / suits (initiate a proceeding)
2. Counter-affidavits & replies
3. Written submissions / synopses / list-of-dates (post-hearing road-maps of arguments)
4. Compilations (authorities, case-law bundles)
5. Paper-books (complete record: pleadings, evidence, orders)
6. Misc. applications (interim reliefs, amendments, condonation, etc.)

All six are now posted digitally—either:

• Automatically by court registry (e-filing portals, 3-PDF paper-books, scanned record-room dumps) or
• Voluntarily by parties (Indian Kanoon, Bar & Bench, SpicyIP, NGO litigation pages).

The sections that follow explain where to obtain each artefact, access requirements, scraping tips, recommended preprocessing and licence notes.

1. Supreme Court of India – richest single trove of arguments

1.1 “3-PDF” Paper-Books (live matters)

Feature	Details	Access route	Licensing
Content	PDF-Main (pleadings & impugned orders) + PDF-Additional (counter-affidavits, rejoinders, IA’s) + PDF-ROP/OR (Record-of-Proceedings & Office Reports)	Available to every Advocate-on-Record (AOR) in e-Filing 2.0 dashboard; download as ZIP via “Paper Book (3-PDF)” link	© SCI OGL-v2 by default; free reuse with citation[1][2]
Coverage	All matters listed before Court No. 1 since Jan 2024 (full roll-out expected 2025)	Need AOR credentials → partner programme or employ in-house AOR	OGL v2[3]

Scraping workflow

1. Script a headless login that picks up fresh ZIPs daily (cookies + OTP).
2. Extract PDFs; run OCR (Tesseract 5 with --psm 1) because some annexures are scans.
3. Chunk via bookmarks (/Title) into logical segments: Facts, Questions, Grounds, Prayer.
4. Store in object storage (e.g., S3) with metadata: item_id, diary_no, party_role, doc_type, page_span.

1.2 Record-Room Scans (disposed matters)

Notice	What it means for you
2015 Public Notice on permanent preservation of Part-I case-records[4][5]	Every SC civil appeal (1986-1993 scanned to date) can be ordered in certified e-form; registry is scanning forward chronologically.
How to order	File an Inspection / Copying Application under Order XII r.2 (SC Rules 2013). AOR receives a DVD or FTP link.
Volume	~500,000 PDFs (~110 TB); includes complete paper-books with counsel drafts, trial pleadings, evidence lists.

1.3 Written-Submission PDFs on main site

Many benches now direct “file concisely bookmarked written submissions within 7 days” and registry uploads them public-facing[6][7].
Query pattern:

https://api.sci.gov.in/supremecourt/{yyyy}/{diary_no}/{diary_no}_{yyyy}_{X}_{YYY}_{ZZZ}_Order_{dd-MMM-yyyy}.pdf

Use simple keyword filters: "written submissions" filetype:pdf site:api.sci.gov.in.

2. High Courts & Tribunals

2.1 e-Filing Portals (nine HCs, all NCLT benches)

• Deliver docket PDFs for petitions, counter-affidavits, rejoinders once Scrutiny Cleared.
• Login via advocate credentials; portal URLs: filing.ecourts.gov.in/.
• JSON API endpoint /efiling/viewDocument?docId={} returns base-64 PDF.

2.2 Cause-List Paste-ins & “Case-Documents” Tabs

• Bombay HC & Delhi HC expose “Case Documents” in HTML—including written submissions.
• Scrape with Selenium, throttled to 1 req/3 s to avoid blocking.

2.3 NGT, SAT, NCLAT

Orders include annexed applicant’s synopsis; paper-books available via RTI under S.4 (1)(b) proactive disclosure.

3. Voluntary Open-Source Repositories of Pleadings

Host	Type of document	Highlights	Licence
Indian Kanoon “Petitions” tab	PDF of filed petitions, counter-affidavits in PILs	35,000+ Supreme Court filings back to 2012	CC-BY-SA[8]
Bar & Bench / LiveLaw / SC Observer	Full‐text petitions & counter-affidavits in constitutional matters	Often uploaded same-day for breaking litigation[9][10]	Public domain (filed in court)
SpicyIP Litigation Database	IPR writs, plaints & written statements	3,800+ PDFs from Delhi HC, Madras HC[11]	CC-BY
NGOs (PUCL, CHRI, Centre for Policy Research)	PIL paper-books, expert reports	Structured zip archives	OGL or CC

Scrape RSS feeds; push to litig-raw bucket; run nightly dedupe by SHA-256.

4. Standard Draft Templates & Forms (for prompt-engineering examples)

Form	Source	Why useful	Citation
Supreme Court SLP Form-28 & Order XVI Rule 4 skeleton	Official SCI forms page[12][13]	Perfect “boiler-plate” to teach LLM the fixed parts vs. variable sections	47
Model counter-affidavit format	SCI affidavit template[14]	Shows jurat, verification language, serial-paragraph structure	50
Writ-petition & PIL samples	Delhi HC PIL compendium[11]		58
Civil plaint drafting guide	iPleaders article + sample plaint[15][16][17]	For training drafting assistant to auto-populate CPC-compliant plaints	35
“Written-submission” style sheet	Latest Court orders specify font, margin, bookmark rules[8][18]	Allows generation of court-ready PDF	21

5. Acquisition Tactics When Documents Are Not Online

5.1 Certified-Copy / Inspection Rights

• SC Rules 2013 Order XII and CPC Order XI r.6–8 give any party or bona-fide researcher the right to inspect case-records on paying nominal fees.
• File physical or e-Filing application → registry emails scanning cost → pay online → receive sealed PDF (open licence per OGL).

5.2 RTI Act 2005

• High Courts & Tribunals are “public authorities”. Pleadings are “information held”. File RTI seeking soft-copies (ask for CD/pendrive under Rule 4(d) of RTI Fees).

5.3 Collaboration With Advocates-on-Record

• AORs already download paper-books daily; sign data-sharing MoU.
• An AOR can inspect any disposed SC record for ₹100.

6. Cleaning & Structuring Pipeline

1. OCR & Text-Layer

   • Use tesseract -l eng+hin+gur --oem 3 --psm 1.
   • Deskew via ocrmypdf --deskew --fast-web-view.

2. Segmentation

   • Regex on common headings: SYNOPSIS, LIST OF DATES, GROUNDS, PRAYER, ANNEXURE P‐….
   • Store offsets to allow retrieval of a single argument passage.

3. De-duplication

   • Many annexures repeat statutes already in your Acts-corpus. De-duplicate with SimHash Jaccard > 0.85.

4. Metadata Enrichment

   • Use SCI diary API to pull bench, judge, arguments concluded date, decision outcome.
   • Tag each paragraph with party side: PETITIONER_ARGUMENT, RESPONDENT_REBUTTAL.

5. Citation Graph

   • Extract (case_cited, paragraph_no) pairs → build precedent-argument graph for Retrieval-Augmented Generation.

7. Suggested Vector Schema for Arguments DB

Field	Type	Example
doc_id	UUID	SC_2025_SLPCiv_4621_WS_Petr
party_side	Enum	PET/SYNOPSIS
court_level	Enum	SC, DELHC
argument_text	Text (chunk 400 tokens)	--
citations	Array	["(2015)7SCC497","AIR1965SC1150"]
legal_issues	Array	["Article 14","Promissory Estoppel"]
embedding	Vector	from instructor-large-legal
date_filed	Date	2025-04-03
outcome	Enum	ALLOWED/REJECTED/PARTLY

8. Licencing & Ethical Compliance

1. Government of India Open Government Licence v2 covers all materials filed in court records unless copyrighted otherwise—free commercial reuse with attribution[3].
2. Personal data: scrub Aadhaar, phone, medical details to comply with SC Privacy Committee 2023 redaction policy.
3. Keep hash-locked originals to prove authenticity; use redacted text for model ingestion.
4. Cite court & diary number in every reproduced snippet.

9. Rapid-Start Checklist

Week	Milestone	Tools
1	Partner with AOR; obtain e-Filing credentials; set up S3 bucket.	Python 3.11, Playwright
2	Automate 3-PDF download & OCR; ingest two pilot cases.	Tesseract + OCRmyPDF
3	Build Mongo / Postgres schema; generate embeddings; test semantic search.	pgvector / Redis
4	File RTI for 25 historic paper-books; add to corpus.	RTI online portal
5	Fine-tune Llama-3-8B-Instruct on 10,000 paragraph pairs (issue ↔ argument).	HuggingFace TRL
6	Evaluate against manual drafting tasks; iterate prompt.	OpenAI evals

10. Conclusion

High-quality argument data—not just bare judgments—is now abundant in India’s digitised court system. By combining:

• Institutional feeds (Supreme Court 3-PDF paper-books, record-room scans),
• High-Court e-filing downloads,
• Voluntary open repositories (Indian Kanoon, media portals), and
• Targeted RTI / inspection requests,

you can assemble a legally compliant, richly annotated corpus of petitions, affidavits, written submissions and drafting templates. With robust OCR, segmentation and metadata-tagging, this corpus becomes prime fuel for:

• • Training retrieval-augmented LLMs that draft court-ready documents,*
• • Suggesting winning argument structures,*
• • Predicting bench-specific preferences,*

and ultimately delivering a next-generation Indian Legal-Research AI.

Harness these pipelines now, and your system will learn not merely what the courts decided, but how the best advocates persuaded them.

[1] https://efiling.sci.gov.in/uploaded_docs/user_manual/3pdf_user_manual.pdf [2] https://cdnbbsr.s3waas.gov.in/s3ec0490f1f4972d133619a60c30f3559e/uploads/2024/01/2024012785-1.pdf [3] https://cdnbbsr.s3waas.gov.in/s3ec0490f1f4972d133619a60c30f3559e/documents/misc/practice.pdf_0.pdf [4] https://cdnbbsr.s3waas.gov.in/s3ec058844c5f00372df2c3c4ee857c245/uploads/2024/01/2024010135.pdf [5] https://districts.ecourts.gov.in/sites/default/files/Supreme%20court.pdf [6] https://api.sci.gov.in/supremecourt/2023/18499/18499_2023_11_54_48939_Order_08-Dec-2023.pdf [7] https://api.sci.gov.in/supremecourt/2016/41681/41681_2016_8_27_50898_Order_27-Feb-2024.pdf [8] https://api.sci.gov.in/supremecourt/2008/2423/2423_2008_16_102_45923_Order_09-Aug-2023.pdf [9] https://www.scobserver.in/wp-content/uploads/2021/10/Petition_filed_by_Advocate_Reepak_Kansal.pdf [10] https://images.assettype.com/barandbench/import/2018/09/Maharashtra-Govt-Counter-Affidavit-Bhima-Koregaon.pdf [11] https://spicyip.com/wp-content/uploads/2018/01/04.-counter-affidavit-uoi.compressed.pdf [12] https://main.sci.gov.in/pdf/Forms/SCI-SLP%20format.pdf [13] https://main.sci.gov.in/pdf/Forms/slp%20format.pdf [14] https://hctlsc.tripura.gov.in/sites/default/files/download_forms/affidavit_1.pdf [15] https://blog.ipleaders.in/sample-plaint-civil-procedure-code/ [16] https://www.scribd.com/document/398703756/DRAFTING-OF-PLAINT-pdf [17] https://www.studocu.com/in/document/chaudhary-charan-singh-university/llb/draft-file/39637925 [18] https://api.sci.gov.in/officereport/1996/76045/76045_1996_2024-05-14.pdf [19] https://inbrief.nswbar.asn.au/posts/21dc6bda18d1036ec37b45178e141b4e/attachment/Practice%20Note%20SC%20CA%201.pdf [20] https://main.sci.gov.in/php/FAQ/5_6246991526434439183.pdf [21] https://cdnbbsr.s3waas.gov.in/s3ec0490f1f4972d133619a60c30f3559e/uploads/2024/01/2024011587-1.pdf [22] https://lawcat.berkeley.edu/record/1219593 [23] https://cdnbbsr.s3waas.gov.in/s3ec047fcc48d22804dbbe9b66b607d513/uploads/2023/11/2023111326.pdf [24] https://www.hcourt.gov.au/assets/cases/08-Sydney/s126-2022/Stanley-DPPNSW_Res.pdf [25] https://api.sci.gov.in/supremecourt/2022/31008/31008_2022_1_1505_39620_Judgement_07-Nov-2022.pdf [26] https://main.sci.gov.in/officereport/2021/5261/5261_2021_2024-01-29.pdf [27] https://archive.org/details/unitedstatessup04publgoog [28] https://www.legalaid.nsw.gov.au/my-problem-is-about/fines/fines-go-to-court/Going-to-court/your-submissions [29] https://efiling.sci.gov.in/uploaded_docs/user_manual/eFM_2.0_Manual.pdf [30] https://www.sci.gov.in/sci-get-pdf/?diary_no=501672023&type=o&order_date=2025-04-08&from=latest_judgements_order [31] https://www.vitalsource.com/products/the-supreme-court-lawrence-baum-v9781071901731 [32] https://cdnbbsr.s3waas.gov.in/s3ec0490f1f4972d133619a60c30f3559e/uploads/2024/01/2024011765.pdf [33] https://supremecourt.nsw.gov.au/documents/Practice-and-Procedure/Practice-Notes/cca-practice-notes/replaced/2021_05_01_PN_SC_CCA_01_-_General.pdf [34] https://www.sci.gov.in/judgements-case-no/ [35] https://us.sagepub.com/en-us/nam/the-supreme-court-compendium/book244744 [36] https://www.scobserver.in/about/supreme-court-of-india/procedure/ [37] https://www.archives.gov/files/dc-metro/washington/m216.pdf [38] https://www.scribd.com/document/806376403/9781544390109-the-supreme-court-14th-edition-original-pdf-1702344470 [39] https://www.courts.state.md.us/sites/default/files/import/coappeals/pdfs/informalpetitionfillablepdf.pdf [40] https://www.supremecourt.gov/orders/journal/Jnl23.pdf [41] https://www.thelawadvice.com/news/supreme-court-to-provide-all-scanned-paper-books-to-aors [42] https://www.nycourts.gov/LegacyPDFS/courts/11jd/supreme/civilterm/CH-FORMS/verified_petition.pdf [43] https://lawgicalshots.com/how-to-draft-a-plaint-check-sample-plaint/ [44] https://supremecourthistory.org/wp-content/uploads/2024/08/Index-1-48-Final.pdf [45] https://www.cafc.uscourts.gov/wp-content/uploads/RulesProceduresAndForms/FilingResources/Petition_for_Writ_of_Certiorari_-_Information_Sheet.pdf [46] https://www.supremecourt.gov/orders/journal/Jnl17.pdf [47] https://supremecourtbc.ca/sites/default/files/2023-04/Petitions_0.pdf [48] https://api.sci.gov.in/supremecourt/2025/33610/33610_2025_11_46_62205_Order_23-Jun-2025.pdf [49] https://us.sagepub.com/sites/default/files/upm-assets/120077_book_item_120077.pdf [50] https://www.jurist.org/news/wp-content/uploads/sites/4/2023/03/Final-Same-Sex-Marriage-Counter.pdf [51] https://www.lawsenate.com/publications/articles/special-leave-petition-slp.pdf [52] https://api.sci.gov.in/officereport/2023/8387/8387_2023_2025-03-06.pdf [53] https://lawhelpline.in/wp-content/uploads/2024/01/Counter_Affidavits_in_Writ_Petitions.pdf [54] https://api.sci.gov.in/officereport/2023/2012/2012_2023_2024-04-09.pdf [55] https://law.duke.edu/sites/default/files/lib/scotus.pdf [56] https://api.sci.gov.in/supremecourt/2020/9880/9880_2020_35_3_24965_Order_04-Dec-2020.pdf [57] https://www.supremecourt.gov/opinions/datesofdecisions.pdf [58] https://images.assettype.com/barandbench/2020-05/5fb0fe40-41bb-4b3c-9d60-4ff9618c176a/Centre_SLP_against_Orissa_HC_order_on_COVID_19.pdf [59] https://www.scribd.com/document/402551077/2924-20190208153133-sample-format-of-special-leave-petition-before-supreme-court-1553109852577-docx