Verdict-labeled claims, outlet credibility scores, and speaker-attributed debate data — programmatic access to the Verum Signal corpus, under a documented methodology.
{ "claim_text": "The unemployment rate fell to 3.4% in March.", "claim_origin": "outlet_claim", "verdict": "supported", "outlet": { "id": "wire.example.com", "name": "Sample Wire" }, "methodology_version": "v1.6", "evaluated_at": "2026-04-15T14:22:00Z" }
Every record served by the API includes the methodology version under which the verdict was produced. The full methodology is publicly documented and versioned — customers can detect changes and audit conclusions against the framework.
Read the methodology →Every record is structured, versioned against a public methodology, and queryable by outlet, speaker, event, or verdict.
Every factual claim from 127 tracked news outlets, classified as outlet claim or attributed claim, with verdict labels under documented methodology.
{
"claim_text": "The unemployment rate fell to 3.4% in March.",
"claim_origin": "outlet_claim",
"verdict": "supported",
"outlet": { "id": "daily.example.com", "name": "Sample Daily" },
"methodology_version": "v1.6",
"evaluated_at": "2026-04-15T14:22:00Z"
}
Credibility scoring for every outlet we track, with full verdict count breakdowns and methodology version stamps.
{
"id": "wire.example.com",
"name": "Sample Wire",
"score": 78.4,
"tier": "published",
"verdict_counts": {
"supported": 624,
"overstated": 178,
"disputed": 22
},
"methodology_version": "v1.6"
}
Speaker-attributed claims from political debates, with real-time verdict labels and full event context.
{
"claim_text": "We've cut taxes by 30% since taking office.",
"verdict": "overstated",
"speaker": { "name": "Jane Smith", "party": "R" },
"event": {
"name": "Iowa Senate Democratic Primary R2",
"date": "2026-05-14"
},
"methodology_version": "v1.6"
}
See real API responses from the live corpus. Pick an endpoint, pick a target, read the response.
Examples shown are illustrative. Fictional outlet names, speakers, and events are used in place of real-world entities. Live API responses reference the actual Verum Signal corpus.
Build claim-tracking dashboards, monitor competitive coverage, and surface outlet credibility trends to editorial teams.
Study claim propagation, media credibility patterns, and the structure of public discourse with reproducible data.
Track every claim a candidate has made in public debates, with structured speaker attribution and verdict context.
Enhance existing media intelligence pipelines with structured claim data and outlet-level reliability signals.
Same endpoints, same data, same methodology. Pick the volume that fits.
Now accepting design partners. Selected partners receive complimentary access in exchange for feedback. Public pricing finalizes at general availability.
An outlet claim is a factual assertion made by the publication itself — in the headline, lede, or body copy written by the journalist or editorial staff. These claims reflect the outlet's own editorial judgment and count toward outlet scoring.
An attributed claim is a factual claim quoted or paraphrased from a named source: a politician, expert, spokesperson, or other third party. These claims reflect what the source said, not what the outlet asserts. They do not count toward outlet scoring.
This distinction is central to the Verum Signal methodology. Read more →
supported, plausible, corroborated, overstated, disputed, not_supported, not_verifiable, and opinion. The first three indicate the claim is well-supported by evidence. overstated, disputed, and not_supported indicate degrees of weakness. not_verifiable and opinion are excluded from outlet scoring — the first means independent verification was not possible, the second means the claim is a value judgment rather than a factual assertion. Full definitions and scoring weights are documented at /methodology.
We're currently working with a small number of design partners to refine the API. Tell us about your use case and we'll be in touch within 48 hours.