Every share leaves a trail. Veridict follows it.

Veridict is a tool that helps you judge whether a piece of content — a claim, headline, post, or (in future phases) a photo or video — is likely to be credible. Instead of a flat "real" or "fake" label, it gives you a score and a list of specific findings, so you can weigh the evidence yourself.

Paste the text, headline, or link into the input box on the home page or the Check screen, then press "Run check." Veridict analyzes it and shows you a credibility score along with the reasoning behind it.

Today, Veridict checks text, headlines, and pasted post content. Photo and video checking are planned for a later phase — you'll see those tabs in the Check screen, marked as coming soon.

Yes, Veridict is free to use during this phase of development.

No account is required to run a check. Accounts may be introduced later to support saved case history across devices.

Yes — that's the main use case. Paste content in before you forward or repost it, so you can decide whether it's worth sharing further.

The score (0–100) reflects how credible the content appears based on sourcing, claim plausibility, tone, and internal consistency. A higher score means fewer credibility red flags were found — it is not a certainty rating.

Because most content isn't cleanly one or the other. A claim can be poorly sourced but still true, or well-written but fabricated. A binary verdict hides that nuance — a score with visible findings lets you see exactly what drove the result.

Findings are the individual credibility signals Veridict identified — things like missing sourcing, emotionally loaded language, an unverifiable claim, or a logical inconsistency. Each finding is shown separately so you can judge which ones matter to you.

"Fabricated" means Veridict found strong, specific red flags (e.g. an impossible claim, a debunked fact-check match). "Uncertain" means the content isn't obviously false, but lacks enough sourcing or context to verify — it's a gap in evidence, not a determination of falsehood.

Veridict scores based on how trustworthy the content looks — its sourcing, tone, and presentation — not on independently verified real-world truth. A true claim written without sources or in an alarmist tone can still score low, because the presentation itself carries the red flags Veridict is trained to notice.

Veridict checks whether content reads as sincere reporting or as satire/parody before scoring it. Genuine satire is marked as such, with an explanation of why it reads that way, rather than being penalized with misinformation findings.

It's possible. Analysis is generated fresh for each check, so minor wording differences or model variability can occasionally shift the exact score, though the overall verdict should stay consistent for clearly credible or clearly fabricated content.

Yes. Veridict is a research aid, not a fact. It can miss context, misjudge tone, or lack information about very recent events. Always use your own judgment before acting on or sharing based on a result.

Not yet in this phase. Current checks are based on the content's internal characteristics — sourcing, tone, plausibility — rather than live cross-referencing against breaking news or a fact-checker database. That cross-referencing is planned for a later phase.

Not reliably. Like any AI-based tool, its knowledge has a cutoff and it cannot verify facts that happened very recently, so treat scores on breaking news with extra caution.

AI-detection signals carry real error rates, in both directions — genuine content can be flagged, and AI-generated content can slip through. Treat these signals as one input among several, not a definitive test.

Read the individual findings rather than just the score — they explain the reasoning, and you may find the underlying claim is fine even if the presentation triggered a flag (or vice versa). Use your own judgment as the final word.

No. A high score means Veridict found few credibility red flags in how the content is presented and sourced — it is not an independent verification of every fact contained within it.

This depends on final production settings, which are still being finalized. During development and testing, avoid pasting sensitive personal information into a check.

No — checks are private to your own session. Veridict does not publish or display what individual users check.

Not currently. Cross-referencing with independent fact-checkers is a planned feature, and any such integration would only involve the claim being checked, not your identity.

Photo and video checking isn't live yet. Once available, this FAQ will be updated to explain exactly how uploaded media is handled and stored.

Shareable case links are planned so you can send someone a result directly, particularly useful for showing why you're not forwarding something further.

No. Running a check is a private action — the original poster is never notified or contacted as a result of your check.

Not currently — each check runs fresh. Caching identical checks so repeat content resolves instantly is a planned efficiency improvement.

Yes — this web app is built to work fully on phones and tablets as well as desktop browsers, with no separate mobile site needed.

Veridict works in current versions of Chrome, Safari, Firefox, and Edge on both desktop and mobile.

Not publicly yet. A developer API is a possible future addition once the core checking pipeline is stable.

Not at this stage — this web app is the current focus. A native app is a possible future direction.

Each check is analyzed by an AI model in real time; response time can vary slightly depending on current load, similar to any live analysis tool.