Conflictly

At a glance

Conflictly puts a geopolitical event feed beside a globe and adds AI summaries, grouped stories and confidence labels. It is useful for finding leads and tracing a developing topic. Treat the labels as part of the app’s analysis: they do not independently establish that a report is true.

Good fit

Readers who want to move from a broad event list to a particular signal and its context. Expanding a story exposed related signals; opening one showed an explanation, named source domains and a Read source control. That makes it a starting point for further reading rather than a reason to stop checking the reporting.

Consider something else when

You need verified incident data, calibrated probabilities or dependable alerts for operational decisions. A precise-looking confidence percentage needs a demonstrated method behind it. In the trial, new events from other countries appeared while a country search remained active, so the visible filter was not a reliable boundary around incoming items.

Start here

Search one country in Search events, expand a story, then open a signal and inspect its source trail. Clear the search when returning to the broader feed. This filter/detail/clear path was exercised without an account; no new verification, vote, chat, notification subscription or audio briefing was requested.

Cost and requirements

The pricing page lists Free at €0 with a 24-hour feed, Signal at €4.90 monthly with 30 days of history, and Strategic at €9.90 with the available full history, Operator and exports. Paid features were not tried. The privacy policy says AI chat and verification inputs can go to external model providers.

Limits to know

Chat retention and verification retention differ: the policy describes 30-day chat deletion, while verification material stays with its associated record until deleted or replaced. No accuracy, calibration or alert-delivery measurement was made. Important claims still need their original reporting, particularly when summaries compress uncertain or conflicting accounts.

Still unclear

• How are the displayed confidence percentages calibrated against independently checked events?