Evidence, not verdicts.
Paste a claim, a question, or an article. Tru8 breaks it into its checkable parts, gathers evidence from 30+ published sources, and returns a signed, organised evidence record.
We organise; you decide.
Free to try · a record in ~90 seconds · no API needed
Most tools hand you a conclusion.
Claims move faster than anyone can assemble the evidence behind them — and every tool answers with a verdict, a score, or a pile of links. If your work has to stand up to scrutiny, none of those is the working.
elements · supports / challenges / context · states · gaps · receipts · signed manifest
Not a verdict. A structured evidence record.
Every check returns the same inspectable structure — what supports each part of the claim, what challenges it, what’s missing, and what was excluded and why.
Element decomposition
Each claim broken into the 1–5 checkable parts that can actually be examined.
Supports / challenges / context
Every source mapped to the element it bears on, with a one-line reason — for, against, or framing.
Tier × type classification
Primary / reporting / commentary, across data, official, news, analysis, opinion and academic. Classified, never scored.
Echo detection
Sources that merely repeat one original are grouped — not counted as independent corroboration.
Gaps, named
What the record could not establish, stated as a finding — not hidden behind a score.
Exclusion receipts
Every source set aside carries a reason. No silent curation.
The summary, then the lenses.
Every check opens on a plain-English summary of what was found. Behind it, the same evidence arranged six ways. Click any screen to look closer.
The whole record, at a glance.
What was examined, how the evidence stacks up on each side, where the sourcing runs thin, and what's missing — before you open a single source.
Every source classified.
Tier × type heatmap and ledger. Filter, sort, browse. Receipts for everything excluded — no hidden curation.
The shape of the conversation.
A citation cascade. See where sources agree, where they diverge, and which are just echoing the same original.
When did each piece of evidence appear?
A timeline of every source, ordered by publication date. See how the conversation developed and where the reporting clusters.
One submission, a retainable record.
Submit a claim, question or URL
Paste a claim, a question, or an article — you choose which of its claims to pursue. Console or API.
Decompose into elements
Each claim is broken into 1–5 checkable factual elements, so evidence maps to specifics, not vibes.
Retrieve external sources at depth
Tru8 searches official data, research, legislation, and reporting — each source classified by tier and type.
Return a structured record
Supports, challenges, context, gaps, receipts, and a signed manifest. You decide what ships.
An honest record has edges.
Not a verdict.
Tru8 organises the evidence and shows how it relates to each claim. It does not score truth or tell you what to conclude — you read the record and decide.
Bounded by what's public.
Evidence comes from external published sources — official data, research, legislation, and reporting — not private records, leaks, or primary fieldwork.
Best on a focused set of claims.
Depth per claim beats breadth. A handful of claims gets the fullest record; very long inputs are decomposed and ranked.
A snapshot in time.
The record reflects what was retrievable when you ran it. Re-running later can surface sources that did not exist before.
The same record,
structured for agents.
Everything above ships as JSON — one call, a signed evidence landscape your agent can act on. Your agent decides what matters.
What Tru8 does — and what it deliberately does not.
- What is Tru8?
- Tru8 is evidence research infrastructure. You submit a URL or a claim, and Tru8 decomposes it into checkable elements, retrieves evidence from more than 30 published sources, and organises the results into a structured evidence landscape — what supports each element, what challenges it, and what is missing. Tru8 does not issue a verdict. We organise; you decide.
- How is Tru8 different from a fact-checker?
- A fact-checker hands you a rating — true, false, misleading. Tru8 hands you the organised evidence and its provenance, and leaves the judgement to you. You see every source, how it relates to the claim, and what could not be found, so you can show your working.
- Who is Tru8 for?
- Researchers, journalists, analysts and developers who need to show their working — people who would rather present an organised evidence trail than ask others to trust a single score.
- What sources does Tru8 search?
- More than 30, across government, legislation, academic literature, economic data, news, archives and more — for example GOV.UK, Hansard, FRED, Semantic Scholar, PubMed, the World Bank and the Internet Archive. Every source that is included, and every one that is excluded, carries a receipt.
- Does Tru8 use AI to decide what is true?
- No. AI is used to decompose claims, retrieve evidence and classify it by source type and tier. It never scores credibility and never issues a verdict. The judgement stays with you.
See the record for your claim.
Paste a headline, article, or claim and read the evidence for and against — in your browser. We organise; you decide.