Intelligence Briefing
Ask a question. Get a source-grounded answer with citations.
strong confidence
10 sources
extractive
Based on 10 verified sources covering Myanmar:
By ZARNI MANN / THE IRRAWADDY Friday, February 24, 2012 [4]
By DAVID STRINGER / AP WRITER / LONDON Thursday, June 19, 2008 [9]
[1]
MM
www2.irrawaddy.com
· 56% match
[2]
MM
www2.irrawaddy.com
· 55% match
[3]
MM
www2.irrawaddy.com
· 53% match
[4]
MM
www2.irrawaddy.com
· 53% match
By ZARNI MANN / THE IRRAWADDY Friday, February 24, 2012
[5]
MM
www2.irrawaddy.com
· 52% match
[6]
MM
www2.irrawaddy.com
· 51% match
[7]
MM
www2.irrawaddy.com
· 51% match
[8]
MM
www2.irrawaddy.com
· 50% match
[9]
MM
www2.irrawaddy.com
· 50% match
By DAVID STRINGER / AP WRITER / LONDON Thursday, June 19, 2008
[10]
MM
www2.irrawaddy.com
· 50% match
How this works
The briefing engine uses semantic search (pgvector embeddings) to find the most relevant articles in the corpus for your question, then extracts key passages with numbered citations. No generative AI is used — all text comes directly from published sources.
Confidence levels
Strong — 5+ relevant sources with high similarity (>50%). The corpus has substantial coverage.
Moderate — 3-4 relevant sources or moderate similarity. Coverage exists but may be incomplete.
Weak — 1-2 sources or low similarity. Evidence is limited — verify independently.
Insufficient — No relevant articles found in the monitored corpus.
Limitations
Briefings reflect only what is in the monitored corpus (686,996 articles). Coverage varies by country and topic. Recent events may not yet be indexed. This is extractive synthesis, not analysis — it shows what sources say, not what to conclude.