PlaigLab — Forensic Document Analysis

Submit evidence

Drag a PDF, DOCX or TXT here, or click to browse.
The document will be investigated against world literature in real time.

DEEP HISTORY SWEEP — walk the citation graph & the topic's all-time canon, plus the web/books tier (Wikipedia, Gutenberg, Internet Archive), scrape up to 14 full texts (slower, exhaustive) MAX — READ EVERYTHING — open & read every open-access candidate's FULL text (not just a shortlist) and run the AI ensemble on each source paper too. Most exhaustive; ~5–10 min. Paywalled bodies stay abstract + shadow-text (legal).

searches OPENALEX · ARXIV · CROSSREF · SEMANTIC SCHOLAR · WIKIPEDIA · GUTENBERG · INTERNET ARCHIVE + deep-scrapes full texts

Investigation in progress

Extraction & parsing
Deep understanding
World-literature search
Citation-graph & canon sweep
Deep PDF scraping
Multi-level comparison
AI authorship scan
Report assembly

FINDINGS —

SIMILARITY INDEX

0%

AI AUTHORSHIP SIGNAL

0%

Exhibits — matched sources

Manuscript heatmap

verbatim copy (5-gram proof) strong match ≥ 72% partial / paraphrase 55–72% semantic match (meaning-level) no source found

AI authorship — section strip

AI detector ensemble

COVERAGE — everything that was searched, screened & scraped

download full manifest JSON ↓

Search warrants issued (queries used)

download PDF report ↓

Plaig·Lab

Investigation in progress

Exhibits — matched sources

Near misses — below threshold, shown so nothing is hidden

Manuscript heatmap

AI authorship — section strip

AI detector ensemble

Style seam map — where the writing style changes