An invitation

The peacebuilding field has more evidence than ever, and too little of it reaches the decisions that matter.

Peace Intel aims to change that.

A decision-support platform built on the peacebuilding evidence base, with a research agent that finds, compares and applies what the literature already knows. Every claim traceable to the passage it came from.

Request access See it work

A project of Effective Peacebuilding Initiative

Introducing Peace Intel

An evidence library, made answerable.

Over 5,000 curated papers, evaluations and policy analyses, fully parsed and semantically processed, read by a research system the way a researcher would read them. Ask a real question. Get a traceable answer.

What it holds

5,000+ curated sources.

Papers, field evaluations and policy analysis, fully parsed rather than catalogued by abstract.

How it answers

Answers, not documents.

It follows citation chains, compares findings across studies, and connects a question to the most relevant cases.

Why it is trusted

Named, cited, traceable.

Every claim links to the study behind it and every citation points to the exact passage. No black-box assertions.

See it work

A question, and where the answer came from.

Beta

You

What happened in Darfur between 2003 and 2011? Who intervened, what did they do, and what resulted?

Peace Intel

Three interventions, three different failures.

Three distinct efforts, worth separating because they failed for different reasons. The African Union deployed AMIS in 2004 with a monitoring mandate it was never resourced to fulfil ●1. Its successor UNAMID was the largest peacekeeping operation in the world at its peak, yet its civilian protection record remains contested ◐2.

A frequently cited displacement figure does not trace to a primary source in this corpus ↗3, so treat it with caution.

How to read a citation

A gradient of trust, stated visually.

Not every source is equally strong, and a system that hides that is asking to be trusted rather than earning it. Every citation carries its tier on its face.

●1

Ingested

The full text is in the corpus. The citation points at the passage, and you can read the surrounding argument.

◐2

Abstract

The abstract is held and the full text is not. Enough to place a finding, not enough to verify how it was reached.

↗3

Web

Found outside the corpus. Deliberately carries no accent colour at all, because it has not been through the same scrutiny.

The tell

Solid, hollow, dashed. The shape carries the meaning, so the tiers survive a colourblind reader, a greyscale print and any future change to the palette.

Who it is for

Two kinds of decision.

Now in pilot Funders

Where the evidence already points.

Strategic briefings, portfolio evidence reviews, proposal due diligence, and the gaps worth funding next.

Implementers

What has worked in a context like this one.

For senior leadership, programme managers and practitioners: logic chains, context-matched sources, and briefs you can hand to a funder.

The corpus

Decades of evidence, in one queryable place.

Assembled and curated by the Effective Peacebuilding Initiative and the Alliance for Peacebuilding, and growing.

5,000+Curated sources

190Conflicts mapped

4.2MCitations indexed

1997Earliest record

Join Peace Intel

Get early access.

The private beta opens in January 2027. Early access includes the full research findings and a say in what gets built.

Just want to be told when it launches? Keep me posted →