PUBLIC-RECORD INTELLIGENCE

Claims are easy.
The record is harder.

ClaimWatch discovers public speeches, videos and reporting, extracts checkable claims, remembers election commitments, and builds evidence packets from inspectable sources.

Hourly discovery 2014 · 2019 · 2024 manifestos Primary-source-first ₹0 core infrastructure

Sources discovered — Loading feed

Promise candidates — 2014 · 2019 · 2024

Promises audited — Outcome evidence checked

Proven missed deadlines 0 Direct comparable evidence only

Claim verdicts 0 Supported 0 · Contradicted 0

PROMISE OUTCOME AUDIT

What the evidence proves about dated commitments

The bar below counts only audited promise records. A passed deadline is shown separately from a proven non-fulfilment: ClaimWatch does not call a promise unfulfilled merely because it cannot find proof.

Audit coverage Loading…

Fulfilled by deadline 0 Deterministic target/result comparison at or before deadline

Target reached · timing unresolved 0 Outcome reached, but evidence does not prove it was on time

Proven not fulfilled by deadline 0 Direct comparable post-deadline evidence falls short of target

Progress documented 0 Implementation activity exists; promised outcome not established

Deadline passed · unresolved 0 Overdue, but evidence is not enough for a final status

Other / insufficient 0 Future deadlines, insufficient evidence, or non-measurable commitments

AUTONOMOUS DISCOVERY

What ClaimWatch is seeing now

Recent official speeches, public videos and reporting. Discovery is not a verdict; it is the queue from which claims enter verification.

Last scan —

Loading source intelligence…

EVIDENCE ENGINE

Claims moving through verification

Each packet keeps the original statement separate from the evidence found for it. Automatic signals help triage; they do not become a published verdict by themselves.

Primary-source candidate Reporting Prior fact-check

Building evidence view…

REPEATED CLAIM MEMORY

One proposition. Every occurrence.

ClaimWatch groups materially equivalent statements into semantic families using multilingual entailment plus numeric compatibility. Repetition is counted separately from the evidence verdict.

Building claim memory…

PROMISE ATLAS

2014 → 2019 → 2024

Machine-extracted commitment candidates from BJP manifestos, with page provenance, measurable targets, deadlines, cross-election similarity and current evidence leads.

Corpus status Loading…

Measurable only

Loading manifesto corpus…

LIVE TAB TRANSCRIPTION · LOCAL AI

Listen to a rally or TV stream and check claims as they are spoken.

Open the speech or YouTube video in a browser tab, click Start, choose that tab in Chrome/Edge and enable Share tab audio. Whisper runs in your browser using Transformers.js; ClaimWatch does not upload the captured audio to its own server. Newly detected factual claims are sent automatically to the evidence checker, and the live tally separates supported, contradicted and still-pending checks.

Whisper · browser inference

Language hint

Idle. The first run downloads a small Whisper model and can take a while.

Best on Chromium with WebGPU. CPU/WASM is used as a fallback. Each audio segment is transcribed locally after capture.

Live transcript 0 segments

Transcript will appear here…

Checkable claims detected from the live transcript 0 candidates

LOCAL TEST BENCH

Drop in a transcript yourself

This first-pass extractor runs in your browser and highlights statements that look objectively checkable. It does not assign truth from wording alone.

No upload required

METHODOLOGY

Fast retrieval. Slow conclusions.

The system is designed to be aggressive about finding evidence and conservative about declaring what that evidence proves.

01 Preserve the claim

Keep the exact source, date, speaker and video/page location before interpretation begins.

02 Prefer primary evidence

Official statistics, legislation, audit reports, parliamentary records and original manifestos come before commentary.

03 Check comparability

Dates, denominators, geography, definitions, units and methodology must line up before numbers are compared.

04 Attribute outside ratings

Prior fact-checks are useful evidence leads; their ratings remain attributed to their publishers rather than silently becoming ours.

05 Keep uncertainty visible

Pending, insufficient and not-checkable are valid outcomes. The system does not manufacture certainty to fill a dashboard.

06 Do not infer intent

A factual contradiction can be documented. Whether someone knowingly deceived is a different claim requiring different evidence.