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

Feature · AI Search · Primary sources · Streaming citations

An AI search engine built on real sources.

Model Diplomat AI Search is the core of the product. Every answer streams from a curated knowledge base of UN documents, treaty text, government statements, court rulings, and academic sources — and every claim links to the actual document behind it.

200+

Primary sources indexed

193

Countries covered

Streaming

Live answer with citations

Sound familiar?

01

Generalist AI cites things that don't exist

ChatGPT, Claude, and Gemini will confidently produce a paragraph citing UN Resolution 1234/56 that has never existed. Model Diplomat does not. Every citation maps to a real document — because Model Diplomat is built on a curated knowledge base, not a fuzzy memory of the open web.

02

Web search wrappers can't reason about primary sources

Perplexity is excellent at retrieving and summarizing web pages. But the web's coverage of a 2007 UN General Assembly vote is thin — and the secondary blog posts that survive Google's ranking are often the worst sources. Model Diplomat reaches for primary documents first.

03

Real research wants follow-ups

Good research isn't a single query. It's: ask, get an answer with citations, follow a thread, re-ask sharper, drill down. Model Diplomat AI Search is built for that loop — every answer suggests follow-up queries, and conversations preserve context across the thread.

What you get.

Ask anything political or diplomatic

Country positions, treaty text, voting history, conflict context, policy comparisons, historical events. Model Diplomat answers in seconds with citations to verifiable sources.

Citations on every claim

Each factual statement links to a numbered source — UN resolution, treaty article, official statement, peer-reviewed paper. Click the citation to verify.

Streaming answers

Answers stream token-by-token as Model Diplomat reasons through sources — so you see the answer take shape and can stop or redirect early if you're going the wrong direction.

Follow-up suggestions

Every answer ends with 3 suggested follow-ups — sharper, narrower, or adjacent queries that take you deeper into the topic without breaking your flow.

Context-aware threading

Conversations preserve context. Ask about Iran's nuclear program, then ask 'and how does Russia view this?' — Model Diplomat carries the topic forward without you re-stating it.

No fabrication guarantee

Model Diplomat's architecture makes it structurally difficult to invent sources — citations come from a verified knowledge base, not pattern completion. When a question can't be answered from real sources, Model Diplomat says so.

Common questions.

What sources does Model Diplomat actually use?

UN documents (resolutions, debates, reports), treaty text, ICJ and ICC rulings, government ministry statements, IMF/World Bank country reports, peer-reviewed IR and political-science literature, and curated news from established foreign-affairs publications. See the Model Diplomat Methodology page for the full source list.

How is this different from Perplexity or ChatGPT search?

Perplexity searches the open web at query time. ChatGPT mostly searches its training data. Model Diplomat searches a curated political and diplomatic knowledge base — so for political and diplomatic questions, our coverage is deeper, more current on primary sources, and substantially less prone to fabrication. For non-political queries, Model Diplomat is not the right tool.

Is it free to try?

Yes. The free tier gives you AI search with citations, country profiles, and the public knowledge base. Pro ($3/week or $10/month) unlocks unlimited searches and the specialist tools (position paper helper, crisis directive analyzer, debate simulations).

Can I cite Model Diplomat in academic work?

Cite the underlying primary sources Model Diplomat surfaces — that's what your reader (or chair, or examiner) needs to verify. Model Diplomat is the research tool, the same way a library catalog is a tool. The citations themselves should be to the actual documents.

Open Model Diplomat AI Search.

Free to try. Ask a question, see the citations. The fastest way to understand what specialist political AI feels like.

No credit card · Works on any device · Free tier always available

Live example — Model Diplomat AI Search

See it answer a real question.

Every Model Diplomat answer cites real primary sources. Nothing fabricated.

What's the difference between the Trafficking Protocol and the Refugee Convention?

The 1951 Refugee Convention defines a refugee as someone outside their country of nationality with a well-founded fear of persecution on five protected grounds (race, religion, nationality, membership of a particular social group, political opinion). It creates a state obligation of non-refoulement — non-return — under Article 33.1

The 2000 Trafficking Protocol (formally, the Palermo Protocol to the UN Convention against Transnational Organized Crime) is a criminal-justice instrument. It defines trafficking in persons (action + means + purpose), criminalizes trafficking and ancillary conduct, and requires victim assistance — but doesn't create the same asylum-style entitlements.2

Practically, the distinction matters at borders: a trafficked person may simultaneously qualify as a refugee (if facing persecution for a protected ground), but the two frameworks don't automatically map. UNHCR's 2006 guidelines on trafficking and refugee status address the overlap directly.3

Sources

1

Convention Relating to the Status of Refugees (1951) + 1967 Protocol

United Nations

2

Protocol to Prevent, Suppress and Punish Trafficking in Persons (2000)

UN Office on Drugs and Crime

3

Guidelines on International Protection No. 7 — Trafficking

UNHCR

Run this query — or your own — in Model Diplomat AI Search.

Try this query
I was able to finish my position paper in 10 minutes.
Ryaan Khan

Ryaan Khan

MUN Delegate

Where this fits in Model Diplomat

Who uses it

Use cases that depend on it

Talk to founder