NeuroDetectLive

Upload and investigate suspicious media.

One workspace to inspect images, audio, video, and documents, trace every finding to the underlying evidence, and prepare an analyst-reviewed report.

Start investigation Request access

ImagesAudioVideoDocuments

NeuroDetect forensic workspaceCase ND-2026-017 · Image evidence

Analysis statusWaiting for evidence review

01 · Source evidence

Evidence and explainability

evidence-frame-017.pngOriginal preserved · 1536 × 1024

An anonymized sample frame of a speaker seated in a neutral studio — **Original frame**Region markers identify the face boundary (R-01), jacket edge (R-02), and background light transition (R-03).

**Grad-CAM response**Strongest activation is localized around R-01, with a secondary response at R-02.

**Attention rollout**Attention remains concentrated at R-01 and extends toward the light transition at R-03.

R-01Face boundaryR-02Jacket edgeR-03Light transition

9.0out of 10

Unified risk score

High review priority

Multiple visual signals align around marked evidence regions. This score is a review aid, not a forensic conclusion.

High

Technical signal findings

Signal IDs preserved

Localized synthesis traceConcentrated response around the face boundaryS-VIS-04Observed

Boundary consistencyJacket and shoulder edges warrant comparisonS-EDGE-02Review

Metadata contextSource metadata preserved with the evidenceS-META-01Available

Metadata and provenance

Evidence ID: IMG-017
Integrity: Hash recorded
Source: Authorized upload
Model context: Version retained

AI-assisted findings

Evidence-linked

Observation

Grad-CAM and attention rollout both emphasize the face boundary at R-01, consistent with S-VIS-04. A weaker edge response at R-02 aligns with S-EDGE-02.

Uncertainty and contradictory signals

Activation at R-03 may reflect the scene's natural light transition. S-META-01 preserves source context but does not validate the media content.

Recommended analyst checks

Compare R-01 across adjacent source frames, inspect R-02 at native resolution, and verify available provenance outside the model result.

Analyst notes

Review source continuity around R-01 before recording the case outcome.

Final determinationPending analyst determination

Built for investigation teams

Review the evidence behind the result.

Fraud and risk teams
Newsrooms and media verification
Legal and government investigators
Trust and safety operations

Fast verification

Move from intake to a structured review path without reducing the case to a single indicator.

Clear results

Pair each observation with the frame, region, segment, or signal it references.

Reviewable reports

Keep model context, selected evidence, uncertainty, and analyst notes together in the evidence report.

Inside the workspace

Detailed views, an AI summary and a report you can download.

Detailed views

Switch between the original, explainability heatmap and attention views, down to regions, frames and audio segments.

LLM-assisted summary

A plain-language summary of the findings, linked back to the evidence it describes. The analyst makes the final call.

Downloadable forensic report

Export the evidence report with observations, uncertainty, provenance, model context and analyst notes.

Downloadable results

Take the structured results with you for case files, audits or your own tools.

NeuroDetect forensic workspaceCase ND-2026-017 · Image evidence

Analysis statusWaiting for evidence review

01 · Source evidence

Evidence and explainability

evidence-frame-017.pngOriginal preserved · 1536 × 1024

R-01Face boundaryR-02Jacket edgeR-03Light transition

9.0out of 10

Unified risk score

High review priority

Multiple visual signals align around marked evidence regions. This score is a review aid, not a forensic conclusion.

High

Technical signal findings

Signal IDs preserved

Localized synthesis traceConcentrated response around the face boundaryS-VIS-04Observed

Boundary consistencyJacket and shoulder edges warrant comparisonS-EDGE-02Review

Metadata contextSource metadata preserved with the evidenceS-META-01Available

Metadata and provenance

Evidence ID: IMG-017
Integrity: Hash recorded
Source: Authorized upload
Model context: Version retained

AI-assisted findings

Evidence-linked

Observation

Grad-CAM and attention rollout both emphasize the face boundary at R-01, consistent with S-VIS-04. A weaker edge response at R-02 aligns with S-EDGE-02.

Uncertainty and contradictory signals

Activation at R-03 may reflect the scene's natural light transition. S-META-01 preserves source context but does not validate the media content.

Recommended analyst checks

Compare R-01 across adjacent source frames, inspect R-02 at native resolution, and verify available provenance outside the model result.

Analyst notes

Review source continuity around R-01 before recording the case outcome.

Final determinationPending analyst determination

Evidence-linked review

Every finding, linked to its evidence.

Region

Compare the original image with localized explainability and technical image signals.

How NeuroDetect works

From uploaded evidence to a reviewed report.

Step 1

Upload evidence

Open an authorized image, audio file, video, or document in a case workspace.

Step 2

Analyze the media

Process the evidence and preserve the model and version context used for review.

Step 3

Review linked findings

Move from the risk level to the exact regions, frames, segments, and signal IDs behind it.

Step 4

Review the evidence report

See observations, uncertainty, provenance, and analyst notes together before recording a determination.

SDK-first integration

One API.
Every stack.

Investigate in the browser workspace, or bring the same analysis into your own tools with native mobile SDKs, backend SDKs, REST and MCP.

$ npm install @neuraldefend/sdk

import { NeuroVerifyClient } from "@neuraldefend/sdk"
const result = await new NeuroVerifyClient().detectImage("selfie.jpg")