Pramaan प्रमाण
Catch the claim while it is still small.
Catches a false claim early, traces where it came from, and publishes the sourced record on the institution’s own domain.
Candidate narrative N-031Claim: “Solar pump money under Scheme S-12 never reached Block 7”
Provenance
- First appearance
- 05:52, open web post from an account created eleven days earlier
- Amplifiers
- Three accounts within forty minutes
- Spread
- Coordinated Confidence Moderate
- Synthetic media
- One edited clip Confidence Low
Evidence file
- 1Budget line, Scheme S-12, current yearBudgetSource grade A1: reliable source, confirmed
- 2अधिसूचना: ब्लॉक 7 के लिए ₹15 लाख स्वीकृतGazetteSource grade A1: reliable source, confirmed
- 3Purchase orders O-7741, O-7748 and O-7752: twelve solar pumpsOrdersSource grade A2: reliable source, probably true
- 4The claim: a post from an account eleven days oldOpen webSource grade F5: cannot be judged source, improbable
Published record
records.ministry.example/claims/
What the record shows on Scheme S-12 in Block 7
4 documents attached · signed · append-only · corrections are added, never rewritten
Illustration ·synthetic records and dataEvery line cites a graded document
Illustration: The claim, as posted, is not supported: the Block 7 money was sanctioned and ordered.IllustrationThe finding of the synthetic case above.
The budget line, the ₹15 lakh sanction and three purchase orders agree. Delivery is not yet on the record, and two of the payments are held. Flagged at 3% of the claim’s eventual reach.
Polity holds the two payments- Discipline
- Narrative integrity
- For
- Communication offices, ministries, regulators, multilaterals
- Deployment
- By design: Customer infrastructure or national cloudBy designNo vendor data egress. The record is published on the institution’s own domain.
- First engagement
- By design: One ministry or one election cycleBy designA pilot starts with one ministry’s programmes or one election cycle, and one result to defend.
Who had seen the claim by the time the record was out
3%
Flagged 1h 40m after the first post
22%
Record out at 5h, on the institution’s domain
Illustration: Of everyone the claim would reach, 3% had seen it when Pramaan flagged it and 22% by the time the sourced record was out. The other 78% saw it with the record already published.IllustrationShares of the synthetic curve above, not measurements.
Illustration ·synthetic case N-031Share of the claim’s eventual reach
From early warning to a record anyone can check.
Illustration: The same case on its own clock: five hours from the first post to a published record that journalists, fact-checkers and readers can check document by document.IllustrationThe times are those of the synthetic case above.
- 1h 40m
Early detection
By design: Flags claims naming an institution, scheme or officeholder by velocity and network shape, not keyword volume.By designThresholds are tuned against the customer’s own analysts during a pilot. A flag goes to a person; it is never published by itself.
- 2h 30m
Provenance
By design: First appearance, amplifiers, organic or coordinated spread, and any synthetic media, each with a confidence level.By designEvery attribution carries a confidence level and its evidence, or it is not handed over.
- 4h
The evidence file
By design: Budget lines, tenders, gazette notifications and court filings, assembled into a brief where every line links to a document.By designA line without a document behind it does not go into the brief.
- 5h
Publication
By design: The verified record goes to a permanent page on the institution’s own domain, documents attached.By designThe institution publishes on its own domain, after its own review. Alethman writes no messaging of its own.
Put the record out before the claim peaks.
A Pramaan pilot runs inside your own infrastructure, for one ministry or one election cycle, and measures how early a false claim about your programmes is caught, and how soon the sourced record is out.
Request a briefing