For the complete documentation index, see llms.txt.
Skip to main content
Model Diplomat LogoModel Diplomat

AI research tool for international moot court

Cite real precedent.
Win your round.

ICJ jurisprudence, VCLT, ARSIWA, regional courts — for Jessup, Vis, Telders, and every international moot. Cited by paragraph. Never fabricated.

ICJ

Jurisprudence by paragraph

VCLT + ARSIWA

ILC commentaries included

Bluebook + OSCOLA

Pinpoint citations

What memorial week feels like.

01

AI fabricates ICJ cases

Mata v. Avianca made it national news. Generic chatbots cite cases that don't exist — and judges check. For moot, that's an instant zero on the issue. Model Diplomat cites real cases by paragraph, or tells you when it can't find one.

02

Westlaw is expensive, JSTOR is slow

Hours of searching for the one ILC commentary that addresses your issue. The relevant ICJ paragraph buried 80 pages deep. Treaty interpretation under VCLT Article 31 written about across fifteen journals — most of which your library doesn't subscribe to.

03

The compromis is 14 pages of nightmare

Half the work is parsing the facts and figuring out which legal issues actually matter. Then you need precedent on each, on both sides. The compromis doesn't sleep; your team only has six weeks.

04

Memorial week, blank page

You have the cases. You have the arguments. Turning them into 50+ pages of structured legal writing, in Bluebook or OSCOLA format, with proper pleadings — that's the part that eats your weekend.

How it works

Ask. Get the case + paragraph. Verify. Cite.

Model Diplomat surfaces the actual ICJ paragraph, treaty article, or ILC commentary you need — with the citation already formatted. You verify, you cite, you move on to the next issue. No more 80-page case PDFs to skim.

ICJ + ICC + regional court jurisprudence
VCLT, ARSIWA, UN Charter article-by-article
Bluebook and OSCOLA citation formatting
Memorial section drafting with primary citations
Jessup memorial · State responsibility
Find ICJ precedent for state responsibility for non-state actor conduct.

The ICJ articulated the “effective control” test in Nicaragua v. United States [1986] ICJ Rep 14, para. 115, holding that for state responsibility to attach to non-state actor conduct, the state must exercise effective control over the specific operations. Reaffirmed in Bosnia v. Serbia [2007] ICJ Rep 43, paras. 391–407 (genocide attribution standard)…

Nicaragua v. US (1986) para. 115Bosnia v. Serbia (2007) paras. 391–407ARSIWA Art. 8 + commentary

Sources at your fingertips

ICJJurisprudence by case + paragraph
VCLTArticles 31–33 + ILC commentary
ARSIWAArticles 1–27 + ILC commentary
ECHR / IACHR / ECJRegional jurisprudence

Built for moot, not general research.

ICJ jurisprudence by case + paragraph

Find precedent on state responsibility, jurisdiction, treaty interpretation, environmental harm, use of force — by exact case name and paragraph. Never invented.

VCLT, ARSIWA, UN Charter article-by-article

Treaty interpretation under VCLT Articles 31–33. State responsibility under ARSIWA Articles 1–27. UN Charter Articles 2(4), 51, 39, 41–42. ILC commentary integrated.

Regional + ad hoc tribunal precedent

ECHR, IACHR, ECJ, African Court — and ad hoc tribunals (ICTY, ICTR, ECCC, SCSL). Cross-referenced with ICJ practice and treaty law.

Memorial section drafting (cited)

Statement of Jurisdiction, Statement of Facts, Pleadings on the Merits, Prayer for Relief. Sourced skeletons you edit into your team's voice. Not a ghostwriter — a research starting point.

Bluebook + OSCOLA formatting

Pinpoint citations formatted the way moot judges expect. Both Bluebook (US schools) and OSCOLA (UK + international) supported out of the box.

Works for Jessup, Vis, Telders, and more

International law moots: Jessup, Telders, Pictet, Manfred Lachs, Nelson Mandela. Commercial arbitration: Vis, Vis East, FDI Moot. Coverage tuned for each.

Why not just use ChatGPT?

ChatGPT fabricated cases in Mata v. Avianca. For moot, that's an instant zero on the issue. We don't do that.

Feature
Model Diplomat
ChatGPT

ICJ citations

Real cases + paragraph numbers

Often fabricated (Mata v. Avianca era)

VCLT / ARSIWA depth

Article-by-article + ILC commentary

General knowledge only

Bluebook / OSCOLA

Native pinpoint formatting

Inconsistent / hallucinated

Regional courts

ECHR, IACHR, ECJ indexed

General knowledge only

Memorial structure

Knows Jessup / Vis format

Generic legal writing

Common questions.

Does it know my Jessup compromis?

Upload it. Model Diplomat ingests the compromis and surfaces ICJ precedent, VCLT articles, and ARSIWA commentary relevant to each legal issue in the fact pattern. You can ask 'find precedent on state responsibility for non-state actor conduct in this compromis' and get cases with paragraph cites.

Does it cite real ICJ cases or fabricate them?

Real cases, cited by paragraph. Model Diplomat is built specifically not to do what general AIs do — fabricate cases like in Mata v. Avianca. If it can't find a real cite for your point, it tells you so you can keep searching, instead of inventing one.

Bluebook or OSCOLA?

Both. Set your team's preference in settings. Citations are formatted with pinpoint references — case name, court, year, paragraph — to your chosen style.

Will it write my memorial for me?

No — and you wouldn't want it to. Model Diplomat surfaces the right precedent and gives you a cited skeleton for each section (jurisdiction, admissibility, merits, prayer). You write the arguments and the prose. It's a research assistant, not a ghostwriter. Most moots prohibit AI-written submissions anyway.

What about Vis (commercial arbitration)?

Supported. Vis and Vis East differ from public international law moots — the focus is on the CISG, UNIDROIT Principles, and arbitration law (UNCITRAL Model Law, NY Convention). Model Diplomat covers those source families and the relevant case law (ICC, SCC, HKIAC, SIAC awards where publicly available).

How does this differ from Westlaw or LexisNexis?

Westlaw and Lexis are exhaustive legal databases — but they're search engines, not research assistants. Model Diplomat is an AI research layer on top of public-international-law sources, designed to surface the relevant paragraph from the relevant case in seconds instead of hours. Use both: us for fast research, them for verification of the cite.

Is the free tier enough to evaluate this for moot?

It's enough to test if we cite real cases. The free tier gives you 5 cited chats per day — enough to throw your trickiest research questions at it and verify every citation. If we pass that test, the paid tier removes the limit so you can actually do memorial-week research.

Your next memorial.
Actually cited.

5 cited chats/day on the free tier. Throw your trickiest research questions at it. If it passes the citation test, upgrade for memorial week.

No credit card · Cancel anytime · Free tier always available