VerifyX
Truth Beyond The Noise
Truth Agent · Autonomous AI Investigation

Cut through the noise.
Find the verifiable truth.

VerifyX deploys an autonomous AI agent to break down claims, search the open web, weigh source credibility, and deliver an evidence-backed verdict — with full reasoning, not a black box.

Deep Search requires a Tavily key in Settings
7
Verdict Categories
Sources Searched
100%
Reasoning Shown
0
Black Boxes
How It Works

One agent. Full investigation.

Truth Agent doesn't just search — it reasons through the same process a professional fact-checker would.

Sub-Claim Decomposition

Complex claims are broken into atomic, individually-checkable statements before any search begins.

Deep Multi-Source Search

Live web search across news, government, academic, and fact-checking sources — with citations for every claim.

Source Trust Scoring

Every source is weighed by type, track record, and transparency — not all evidence counts equally.

Contradiction Detection

The agent actively looks for evidence that disagrees with the claim, not just evidence that confirms it.

Confidence Scoring

A transparent breakdown of evidence strength, source reliability, and overall confidence — no single opaque number.

Full Investigation Reports

Every verdict comes with a professional report: evidence, sources, timeline, and full reasoning trail.

No active investigation

Submit a claim from the home page, or start one right here.

Media Forensics

Image Investigation

Upload an image for AI-assisted visual analysis and contextual investigation.

VerifyX does not run real deepfake detection, metadata forensics, or reverse image search in this version — those require specialized forensic models this platform does not have access to, and it will not fabricate a score for them. What you'll get below is the AI's visual description and contextual reasoning about plausibility, clearly separated from anything that would need real forensic tooling.

Drop an image here, or click to upload

JPG, PNG, WEBP — analyzed entirely in this conversation, not stored

Dual Investigation Mode

Prove It True. Prove It False.

Truth Agent investigates a claim from both directions at once, then weighs the two cases against each other.

Adversarial Mode

Truth Battle

Two AI agents argue opposite sides of a controversial claim. A Judge AI weighs the evidence and declares a winner.

Archive

Past Investigations

No investigations yet

Completed investigations are saved here automatically for this session.

Trust System

How VerifyX Scores Sources

Every piece of evidence is weighed, not just counted. Trust scores combine source type, transparency, and track record.

Scoring Factors

Trust Score (0–100) is calculated from: source type baseline, citation patterns, transparency of authorship and methodology, and consistency with corroborating sources found during the same investigation. No source is ever treated as automatically 100% reliable — including government and academic sources, which can also be wrong, outdated, or biased.

This Session
Saved in this browser tab only
Overview
Bookmarks
Search History
Collections
0
Investigations
0
Bookmarked
0
Collections

Recent Activity

Community

Submitted & Discussed Claims

Vote on completed investigations, discuss findings, and flag suspected misinformation.

No community claims yet

Investigations you complete will appear here for community voting and discussion.

Admin

Platform Dashboard

Investigations
0
This session
Active Sessions
1
This browser tab
Flags
0
Pending review
API Calls
0
This session

System Health

ServiceStatusDetail

Recent Investigations

ClaimVerdictConfidenceTime
About VerifyX

Built for transparent verification

VerifyX is an AI-powered investigation tool. It does not claim infallibility — it claims transparency. Every verdict is accompanied by the evidence and reasoning that produced it, so you can judge the judgment.

Truth Agent runs on a language model connected to live web search. Its accuracy depends entirely on the quality of available sources and the model configured in Settings. Always treat AI-generated verdicts as a strong starting point for your own judgment — not a final authority, especially for high-stakes claims.

VerifyX does not perform real forensic image or video analysis (deepfake detection, metadata forensics, reverse image search) in this version — that requires specialized models VerifyX does not fabricate results for. Where a feature would need that, VerifyX says so plainly instead of inventing a score.