Every capital, on the record.
The retrieval-first AI research engine for foreign policy. We read primary sources in every relevant language — treaties, MFA statements, UN records, court filings, think-tank PDFs — and return a brief where every claim links to a real document. No fabricated citations.
China Q3 GDP misses forecast as deflation deepens
稳中求进 加快部署新质生产力
Sahel : Paris acte définitivement le retrait
Brussels weighs CBAM expansion to Asian imports
MEA: India’s engagement is multipolar, on principle
Африканский корпус: переброска в Сахель
防衛白書 改定 — 「台湾有事は日本有事」
Council split on de-risking package; vote slips
2M+
DocumentsPrimary documents
1,400
PublicationsPublications, daily
22
LanguagesLanguages, native
Critical foreign-policy sources live in places ChatGPT can’t reach.
The most consequential statements out of Beijing, Moscow, Brussels or Delhi rarely show up in the English-language press the same day — if at all. The signals that move a policy week are posted to MFA websites, Telegram channels, leaked memos, and paywalled trade publications that general LLMs were never built to read.
Non-English primary
MFA statements, white papers, and the policy press of record — published only in the original jurisdiction’s language and rarely translated within the news cycle.
A 国台办 statement on cross-strait policy, three days before it’s paraphrased in English.
Sub-national channels
Provincial MOFA pages, ministry sub-portals, regional press, embassy releases — primary documents that never reach a search engine’s top page.
An MEA permanent-rep statement, posted to a Geneva mission page, never picked up by the wires.
Informal statements
Foreign-minister tweets, leaked council readouts, Telegram channels, post-summit press scrums — where the most candid signal often lands first.
A Russian deputy FM’s Telegram post on Africa Corps redeployment, hours before TASS confirms.
Behind access gates
Paywalled trade publications, PDF-only government releases, login-walled think-tank reports, archives that require institutional access.
A Caixin investigation behind a Chinese paywall, reframing the property-workout narrative.
Every claim, a chip.
Inline citation chips on every paragraph. Click any chip to see the underlying MFA statement, treaty draft, or think-tank paper — opened directly to the cited passage.
India’s voting record on UN General Assembly resolutions concerning the Russian Federation, 2022–2025.
Between March 2022 and October 2024, India abstained on each of the seven General Assembly resolutions concerning the situation in Ukraine on which it cast a vote, including the emergency special session resolution of 2 March 2022 and the 14 November 2022 resolution on humanitarian consequences.
The Ministry of External Affairs has consistently called for adherence to the Charter and for respect for the territorial integrity of states, without naming the Russian Federation as the author of the conduct at issue. The Permanent Representative’s explanation of vote framed the abstention as a continuation of India’s long-standing position favouring diplomacy and dialogue.
The bilateral track has, over the same period, expanded rather than contracted. The 9 December 2024 India–Russia joint statement reaffirmed the “special and privileged strategic partnership” without revisiting the question of Ukraine. The framing is consistent with the literature on non-alignment as updated for a multipolar order.
Built for people who can’t afford the wrong read.
If your work depends on knowing what a capital is actually saying — not what an English-language wire paraphrased two days later — this is for you.
Foreign-affairs journalists
Source from the original record before it’s repackaged. Catch the ministerial statement that didn’t make the wire.
Reporters · Editors · Foreign desks
Policy researchers
Surface the primary documents your paper needs to cite, in the language they were issued in.
Fellows · Postdocs · Research associates
Government & consular desks
Read what every relevant capital is saying about your file, in parallel, with citations your team can defend.
Desk officers · Policy units · Embassy staff
Boutique consultancies
Brief clients with multi-capital reads built from primary sources, not consensus paraphrase.
Partners · Analysts · Briefing teams
Also where we started · 86,000+ students
Free tier on every workflow
The same retrieval engine, built for competitive academic work.
Position papers, evidence cards, Jessup memorials, Mock Trial direct and cross, IB Extended Essays, AP US Government FRQs, Lincoln-Douglas value premises, annotated bibliographies. Every claim is cited to a real document, exportable in MLA, Chicago, APA, or footnotes — so it stands up to a chair, a judge, or a teacher.
Model UN
Position papers · crisis notes · directives
Mock Trial
Openings · direct · cross · objections
Jessup / Moot
Memorials · jurisdiction · oral pleadings
IB Diploma
Extended Essay · History IA · Paper 2
AP / NHD
DBQs · FRQs · annotated bibliography
Speech & Debate
LD · PF · Policy · WSDC · Congress
How a brief gets made.
Retrieval first, synthesis second. Each step is auditable, and every claim traces back to a document you can open in one click.
Step 01
Retrieve
Multi-agent crawl across MFA portals, the UN system (UN, ICJ, OECD, IMF, World Bank, EU institutions), think tanks, and 1,400 publications in 22 languages — including paywalled trade press and PDF-only releases.
Step 02
Verify
Each candidate source is fetched in full, parsed page-by-page (not snippet-and-title), and resolved through official redirects. If a document doesn’t resolve, it doesn’t ship.
Step 03
Synthesize
Capital-by-capital reads are composed into one brief. Inline citation chips anchor every claim. Conflicting positions surface side-by-side — no consensus paraphrase.
Step 04
Export
Open documents at the cited paragraph, push to Zotero, or export footnotes in MLA, Chicago, APA, or your house style. Schedule the same query on a cadence and receive briefings by email.
Why not just use ChatGPT, Perplexity, or Elicit?
Each is excellent at what it’s built for. None of them are built for foreign policy with verifiable citations in the original language. Here’s the gap.
Three ways to use it.
Same indexing and synthesis layer — you choose how it shows up in your day.
Research
Ask anything
Type a question; multi-capital synthesis comes back with a primary citation on every claim, in your language. The primary surface — and the heart of everything else.
“What does Delhi actually want from BRICS?”
DELHI
EN/HI
Strategic autonomy.
BEIJING
ZH
BRICS+, on its terms.
MOSCOW
RU
Anti-Western bloc.
BRASÍLIA
PT
Trade-first, sovereign.
Briefings
Standing queries, delivered.
Set a question on a cadence — daily, weekly, before each committee. We re-run it and email the brief.
Courses
Catch up on a region
Hundreds of structured courses across treaty law, regional history, economic statecraft, and committee procedure. Sometimes you need the background before the brief.
Intermediate
Cross-strait dynamics
12 lessons
Advanced
Sahel security architecture
9 lessons
Foundations
EU trade-defence toolkit
8 lessons
Intermediate
BRICS+ since Johannesburg
11 lessons
FAQ
How is this different from ChatGPT, Claude, or Perplexity?+
Those retrieve and weight English-language sources by default. Model Diplomat is retrieval-first across 1,400 publications in 22 languages, weighted by readership inside each capital. Every claim is anchored to a primary document — MFA statement, UN record, ICJ filing, treaty text, peer-reviewed paper — that you can open at the cited paragraph. The output is multipolar by construction, not by prompt.
How is this different from Elicit, Consensus, or NotebookLM?+
Those are built for journal articles (Elicit, Consensus) or your own uploaded documents (NotebookLM). Model Diplomat is built for the political record: MFA portals, the UN system, ICJ, OECD, IMF, World Bank, EU institutions, government releases, think-tank PDFs, and the foreign-language policy press. Different corpus, different workflow, same retrieval-first discipline.
Are the citations real?+
Yes. Every claim links to a primary source. No fabricated footnotes. If a claim has no source, it doesn’t ship. We resolve official-publication redirects, parse PDFs in full, and verify each citation before it’s rendered.
Can I export citations in MLA, Chicago, or APA?+
Yes. Every brief exports as MLA, Chicago, APA, or footnotes — and pushes directly to Zotero. The cite-format that wins your chair, judge, teacher, or editor is one click away.
What languages do you cover?+
22 languages, read natively — including Chinese, Russian, Arabic, French, German, Japanese, Hindi, Portuguese, Spanish, Korean, Turkish, and Persian. We index in the original and brief you in yours.
Is this only for foreign-policy analysts, or also for students?+
Both. 86,000+ students use Model Diplomat for Model UN position papers, Mock Trial direct and cross, Jessup memorials, IB Extended Essays, AP US Government FRQs, NHD annotated bibliographies, and Lincoln-Douglas value premises. The same retrieval engine — verifiable citations that hold up under scrutiny — works for analysts, journalists, postdocs, and policy desks.
What does it cost?+
You can get started for free — no signup on your first query. The free tier covers daily search, country briefs, and learning modules. For deep research runs, scheduled briefings, persistent memory, and team workspaces, check out our pricing.
Walk into the meeting with what each capital actually said.
Book a 30-minute call. We’ll walk through the regions you cover, show you a real brief from this week, and set up a trial tailored to your beat. Every citation resolves. Every capital is on the record.