One claim appears
A recycled safety warning returns with a new city name. One submission is a signal—not yet an outbreak.
- Observed
- 1 thresholded regional signals
- Mutation
- 1 distinguishable message variants
The Yaani Labs vision
From checking one forward to understanding an information ecosystem.
WhatsApp is one of the world's largest conduits for ideas, warnings, rumours and persuasion—yet the public cannot see how those narratives emerge, mutate or move. Yaani begins with a useful answer for one person and compounds those checks into shared intelligence.
message→case→lineage→weather→intervention
A scroll-driven future prototype
Watch a narrative become weather
Yaani narrative weatherDay 1Scroll to advance ↓
Observed regions1
Linked variants1
Current stateOne claim appears
Language variants emerge
The warning is translated, its location changes and the deadline becomes more urgent. The factual core stays recognisable.
- Observed
- 3 thresholded regional signals
- Mutation
- 3 distinguishable message variants
A news event gives it fuel
Real storm coverage is attached to the old message. New context increases plausibility and accelerates mutation.
- Observed
- 6 thresholded regional signals
- Mutation
- 6 distinguishable message variants
Corrections join the system
The living case updates, sibling versions connect and portable tells travel back out without exposing private people.
- Observed
- 8 thresholded regional signals
- Mutation
- 8 distinguishable message variants
Privacy boundary Voluntary submissions and public sources can contribute aggregate signals. Private people and sparse locations never become dots on this map.
A compounding roadmap
Every phase leaves a better product behind.
If work stopped after any phase, the public checker would still be more accurate, more explanatory and more useful than before.- 01Live now
A fact-checker that teaches memetics
Check text and media, inspect the evidence, understand the spreading pattern, and send a useful answer back to the chat.
- 02Built
A compendium beneath every check
All 29,447 acquired fact-check articles now feed 24,016 searchable claim families in the same public memory as Yaani's current checks. Known facts, evidence gaps, memetic mechanisms and 72,026 mutation hypotheses can guide a new check without turning historical analysis into a current verdict.
- 03Live now
Living cases and an emergence desk
Recognise sibling claims, show changed wording without inheriting an old verdict, follow evidence updates, and watch approved public sources enter a clearly labelled research field.
- 04Live, gated
Narrative weather for India
Council-reviewed signals can enter Pulse now. Atlas, mutation lineage and news context publish only when their separate independence, privacy and evidence gates clear; an honest empty map is better than invented spread.
- 05Partner-ready
A bounded newsroom and research desk
Scoped intake, privacy-thresholded exports, pull alerts, credential limits and a de-identified operations console are ready for real integrations without exposing contributors or raw forwards.
- 06Next
An intervention lab
Test corrections, response formats and portable tells to learn what helps communities resist the next mutation—not merely win an argument about the last one.
The questions this lab can answer
Not just “is it false?”
- What is emerging?Spot a new narrative before a professional debunk already exists.
- What changed?Compare wording, images, targets, urgency and authority costume across variants.
- Why now?Connect recurrence to relevant news, history and social context without confusing correlation with cause.
- Where is it moving?Show thresholded regional and language signals without exposing private people or tiny groups.
- What did it cause?Separate documented consequences from inference and speculation.
- What helps?Measure whether a correction, explanation or portable tell survives contact with the next mutation.
Ambitious by design. Careful by default.
An observatory must not become a surveillance system.
Public maps require minimum independent-contributor thresholds. Private submissions are evidence for analysis, not dots on a public map. Observed activity, reconstructed lineage and causal inference must remain visibly different.
The prototype above is intentionally illustrative. Yaani will only label a surface live when the underlying collection, review, privacy and correction systems are genuinely operational.
The vision begins with one useful check