ChatGPT–Capterra Evidence Auditor

Check whether a source-audit row cleanly separates a Capterra mention, citation, SaaS recommendation, and measurable ChatGPT referral. The checks flag internal contradictions and causal overclaims; they do not inspect ChatGPT or Capterra.

Private by design: parsing and checks run only in this browser tab. No upload, network request, cookie, analytics call, or browser storage.

1. Load the 43-field ledger

Delete every EXAMPLE-REMOVE row before operational reporting. Do not paste account identifiers, private analytics details, raw cookies, confidential prompts, or personal data.

2. Evidence summary

0answer rows
0example rows
0findings
0claim-ready rows

3. Findings

Load a ledger and run the audit.

What the checks mean

CheckPass conditionBoundary
RecommendationA recommended product has a name, position, rationale, and non-zero count.Does not identify the source that caused selection.
Capterra citationA citation has a Capterra URL, page type, role, supported claim, and count.Does not prove general reliance or causal influence.
ReferralAn observed analytics session has source/medium; ChatGPT links normally retain utm_source=chatgpt.com.Does not connect the visit to an unseen source.
Neutral rateSource-requested prompts are retained but excluded from neutral-source frequency.Instruction compliance is a different measurement.
ReproducibilityPrompt and answer hashes, UTC time, capture path, and context are preserved.A hash does not validate factual accuracy.

This utility validates ledger consistency only. It does not authenticate captures, open cited URLs, query an AI service, read analytics, or establish hidden retrieval and causality.