Which story are we checking?

We do not save your API key. It is sent with each request and held only in this page's memory. Use a temporary key with a spend limit and revoke it afterwards.

Experts appear as “Expert #n”.
Ties found in the research record. A missing tie is not proof.

Case graph

About TieTrace

Case initiation

Let’s follow the evidence.

Check how connected an expert quoted in a story is to a company, using documented public research records. Every tie reveals its DOI, and every “no tie found” reveals its searched scope.

Investigation standard & epistemic boundaries

What a result establishes

  • A tie is a stored edge in the research record, shown with its DOI.
  • A “no tie” names its scope: hops, years and funder coverage.

What it does not prove

  • Intent, bias or editorial influence.
  • That a missing tie means there is none: absence is not proof.

The blindspot

Unseen ties in the quotation box.

When deadline pressure hits, reporters quote authoritative academics. But funding and affiliations are buried in publication records that no one has time to read in full. TieTrace reads them and shows each tie with the paper it rests on.

Typical deadline failure:
  • The expert is quoted from a press release.
  • Funder lines sit at the end of a paper nobody opens.
  • The story publishes with no record context.
Investigative dissectionIllustration
Published article draft“The study found no link,” said Expert #1.

TieTrace detected research ties

Paper funded by the companyShown with its DOI and funder name
Tier A1
Co-author of a Tier A personNetwork context only, not validated
Context

Pipeline

Four steps from text to source record.

Claude Code reads the story and resolves it against the research graph with read-only queries.

Step one

Find the experts

Reads the story and picks the research case for the company and topic.

Step two

Look up in graph

Names resolve to OpenAlex authors. If a name is ambiguous, you pick; it never guesses.

Step three

Trace documented ties

Funders, affiliations and co-authors, each query shown in the live evidence panel.

Step four

Write sourcing note

DOIs for every tie, and the scope for every “no tie found”.

Graph traversal traceUp to 3 hops
authoredfunded bypart of ExpertOpenAlex ID PaperDOI Funderon the paper Companyor a bottler

Direct primary recordCompany link

Provenance hierarchy

Tier A1, A2, B and C.

Every documented connection is sorted into a tier, so newsrooms know what is evidence and what is context.

Tier A1 — directDirect evidence

Funded or affiliated, on a small paper

Authored a small paper (25 authors or fewer) funded by the company, or listed a company affiliation on one.

Shown with DOIs
Tier A2 — directDirect evidence

Grant, bottler, or large paper

Foundation grant, bottler affiliation, or the only evidence is a large multi-author paper.

Shown with DOIs and the reason
Tier B — contextNetwork context only

Recent co-author of a Tier A person

Two or more recent small papers (since 2021) with a Tier A person. Not a finding.

Not validated
Tier C — contextExploratory lead

Indirect collaborator path

Within 2-3 such co-author hops of a Tier A person. Must be flagged as non-direct.

Not validated

Epistemic honesty

Why “no tie found” is never bare.

A missing tie can simply mean missing data. Claiming someone has “no ties” without showing what was queried is dangerous, so TieTrace prints the bounding box of every negative finding.

Absence of a documented record within searched indices is not proof of non-existence.
Audit result statusNon-detected within scope

No tie found in the research record

Target subject: Expert #n

Bounding scope parameters
Graph hop depth
up to 3 co-author hops
Recent years
2021 onwards
Funder coverage
stated per graph and per expert
Unsearched
gifts to universities, unlisted funders

Every real result prints these values from the graph it searched.

Verifiable metrics

What we tested, honestly.

Pass rules were fixed before the runs. Failures are shown with the passes.

E3 known case4 of 4

Publicly confirmed GEBN leaders found as Tier A with DOIs.

E8 golden tests15 of 15

Pass.

E1 placebo lift (Tier B)FAIL 0.96×

Tier B is not a sign of unusual closeness.

E2 temporal hold-outFAIL

The score (AUC 0.716) does not beat a plain count of Tier A co-authors (0.749).

Because E1 and E2 failed, only Tier A is a finding. Full results

Human gateway enforced
1. Name checkAmbiguous names: you pick the author
You pick
2. VerifierEvery claim checked against the graph
Automatic
3. Note approvalA confirmed click, with your name recorded
You approve

No verdict ever auto-publishes

Editorial gatekeeping

Who decides: a journalist approves every note.

The model never records a verdict and never approves a sourcing note. TieTrace provides the evidence trail and puts a human checkpoint between the graph data and your published article.

  • A sourcing note can be approved only after the Verifier checks every claim.
  • Verdicts and approvals need a second, confirming click.
  • The reviewer’s name is recorded with each decision.

Forensic boundaries

Strict ethical boundaries.

We say “ties found in the research record”. We present documented relationships and never allege motive.

What this establishes

  1. Disclosed funding: a funder named on a paper, with its DOI.
  2. Declared affiliations: an affiliation listed on a paper.
  3. Traceability: every tie maps to a stored edge and a DOI.

What it does not prove

  1. Bias or wrongdoing: sponsorship is not proof of altered results.
  2. Editorial influence: ties do not show a funder shaped the conclusions.
  3. Right of reply: findings are leads that need a formal inquiry before publication.

Ready to verify your story?

Paste your quote or article draft below to trace documented ties in the research record.

Nothing is sent until you press Run Check.

Ties found in the research record. Strict non-inference. DOI on every tie.