ResAIKit
Research Integrity Toolkit
Work in progress. ResAIKit is under active development. Features, indicators, and pricing may change. The tool is available for testing and feedback. Do not use for production editorial decisions yet.
Pre-submission and editorial research-integrity screening

Detect research-integrity risks before they become publication problems.

AntiAI helps editors, reviewers, and institutions screen scientific manuscripts for AI-generated text, manipulated figures, and statistical fabrication signals — producing structured evidence for expert review, not black-box accusations.

See all indicators

Designed for triage, documentation, and transparent human decision-making.

Integrity scan report
Manuscript_ClinicalStudy_v4.pdf
Needs review
AI-like text pattern
Elevated
Figure anomaly signals
Moderate
41
Statistical plausibility
High concern
83
Review priorities
Repeated generic transition patterns in discussion section
Two figure panels show unusual texture continuity
Reported p-value distribution inconsistent with sample size
Three-layer screening

From manuscript text to figures and statistical traces.

T

AI-text screening

Identify linguistic patterns, generic framing, weak methodological anchoring, citation drift, and AI-like academic phrasing across 32 indicators.

F

Image & figure integrity

Flag visual anomalies, duplicated regions, suspicious figure construction, and AI-assisted image fabrication indicators across 47 checks.

S

Statistical fabrication signals

Screen distributions, p-values, effect sizes, digit patterns, internal inconsistencies, and improbable reporting structures across 65 indicators.

Workflow

A defensible review trail, not a yes/no verdict.

AntiAI turns manuscript screening into a documented process: upload, analyze, inspect flagged evidence, export a report, and route the case to the appropriate reviewer or integrity officer.

01
Upload
PDF, DOCX, LaTeX, figures, tables, or manuscript sections.
02
Screen
Run text, image, and statistical modules in parallel.
03
Review
Inspect evidence with transparent, indicator-level explanations.
04
Document
Export a structured PDF report for editorial files.
Outputs

Reports built for editorial decisions.

Each report separates indicators from conclusions, highlights passages and figures requiring inspection, and records the reasoning path used in triage.

Evidence-first
Flags are tied to observable fragments, figures, or statistical values.
H
Human-review oriented
Language supports expert assessment rather than automated accusation.
M
Modular
Use text, image, statistics, or combined manuscript screening.

Add research-integrity screening to your editorial workflow.

Start with a sample manuscript, compare the report structure, and define the review thresholds appropriate for your journal or institution.