About This Project

This is a real-time observatory of the information environment surrounding the US-Israeli strikes on Iran that began on February 28, 2026. It is not a news service. It does not break stories, verify claims, or produce original reporting. Its purpose is to monitor how multiple media ecosystems are processing, framing, amplifying, and contesting the same set of events — and to surface the analytical patterns that emerge from reading them together.

What We Monitor

The dashboard ingests content from approximately 55 web sources and 50 Telegram channels spanning multiple information ecosystems:

Russian — state media, military blogs, official channels (~65% of Telegram volume)
Iranian — state broadcasters, foreign ministry, Western-Farsi services
Israeli — broadsheets, OSINT channels, military commentary
OSINT — open-source intelligence aggregators (~15% of Telegram volume)
Chinese — state wire, official media, policy commentary
Arab & Gulf — Al Jazeera, Al Arabiya, resistance-axis media, Gulf state outlets
Turkish — state and semi-official media
South & Southeast Asian — Pakistan, India, Indonesia, Malaysia
US — official statements (CENTCOM, State Department), hawkish policy outlets, wire services
African & other — South Africa, Egypt, Nigeria, Central Asia, Caucasus

This corpus skews heavily toward non-Western and non-English sources by design. The mainstream Anglophone perspective is abundantly available elsewhere. Our value is in juxtaposing ecosystems that rarely appear side by side — reading Russian military bloggers alongside Iranian state television alongside Israeli OSINT alongside Chinese foreign ministry statements — and noting what each ecosystem emphasizes, omits, amplifies, or contests.

Methodology

Editorials are generated at regular intervals using AI-assisted analysis (Claude, by Anthropic). Each editorial cycle works as follows:

All new Telegram messages and web articles since the previous editorial are collected
Irrelevant content (unrelated conflicts, domestic news, spam) is identified and removed
Six analytical perspectives are applied to the same data, each through a distinct professional lens: military operations, great-power dynamics, escalation theory, energy and economic exposure, Iranian domestic politics, and information ecosystem dynamics
A lead editor synthesizes the strongest insights into a single published editorial under one voice

The analyst perspectives are simulated — they are not real individuals. They are a methodological device for ensuring that the same data is examined from multiple disciplinary angles before synthesis, reducing the risk that any single framing dominates the analysis.

Narrative Threads

The dashboard tracks approximately two dozen narrative threads — recurring storylines that appear across multiple sources and ecosystems. These are identified by keyword patterns applied to the full corpus and rebuilt regularly. Each thread includes a timeline of all matching messages and articles, color-coded by ecosystem, allowing readers to trace how a story developed across different information spaces.

Interpretive Cautions

Source composition shapes perception. Because the Telegram corpus is roughly 65% Russian-language, the raw message feed will appear to over-represent Russian perspectives. This is a function of which channels produce the most volume, not an editorial choice. The editorials attempt to weight analytical attention proportionally to significance, not volume.

We report claims, not facts. In a fast-moving conflict with multiple belligerents making contradictory assertions, almost nothing can be independently verified in real time. The editorials attribute all claims to their source ecosystems. When we write that a source "reports" something, we mean that the source made that claim — not that it happened.

We follow the data. The editorials respond to what the sources are actually saying. If a topic is not yet appearing in the media ecosystem — regardless of how analytically important it might seem — we do not introduce it. This is a deliberate methodological constraint: we are observing the information environment, not contributing to it.

AI-assisted analysis has limitations. The editorial synthesis is produced by a language model. It can misread nuance, miss cultural context, or over-weight dramatic claims. The multi-perspective methodology is designed to mitigate these risks, but readers should treat the analysis as a structured starting point for their own assessment, not as a finished intelligence product.

What This Is Not

This is not an intelligence product, a policy recommendation, or a journalistic publication. It is a research instrument — a way to watch how information moves through contested spaces during a crisis. It makes no claims about what is true. It tries to show you what is being said, by whom, to what audience, and how those claims interact across the boundaries that normally separate information ecosystems from one another.

This observatory is maintained as a research project. Source code and methodology are available to collaborators on request.