Probabilistic analysis · human review required

Understand what's human. Detect what's artificial.

TraceMark AI analyzes text, documents and images for signals associated with AI generation — and presents the evidence behind every estimate so your team can make an informed decision.

Inputs
6 formats
Analysis
Sentence-level
Output
Evidence-based

Drag and drop a file to analyze

or click to browse — PDF, DOCX, TXT, JPG/JPEG, PNG

PDFDOCXTXTJPG / JPEGPNGPasted text
Demo

This build ships the complete product architecture and workflow. No detection engine is connected yet, so detector outputs are labelled as demo data rather than scores.

How TraceMark AI works

A documented process, not a single number

Every analysis is designed to be reviewed by a person and defended in a process — from submission to final case decision.

01

Submit content

Paste text or upload PDF, DOCX, TXT, JPG/JPEG or PNG files from any device.

02

Analyze signals

Writing-pattern and metadata/provenance signals are evaluated across the whole submission and sentence by sentence.

03

Review evidence

Likelihood ranges, confidence levels and flagged passages are presented with explanations — never a verdict on authorship.

04

Document the decision

Assign a case status, add reviewer notes, and export a report for your institutional record.

Detectors

Three surfaces, one evidence model

Text, documents and images each contribute different signals. TraceMark AI keeps them separate and explains each one.

Text Detection

Paste any passage and see sentence-level signals, flagged spans and plain-language explanations of what was measured.

Explore text detection

Document Analysis

PDF, DOCX and TXT uploads combine writing-pattern analysis with document metadata and edit-history inspection.

Explore document analysis

Image Analysis

JPG/JPEG and PNG files are checked for generation markers, resampling artifacts and content credentials.

Explore image analysis

Who it's for

Built for institutions that must justify a decision

From a single teacher to a multi-department organization with auditors watching.

Education

Department workspaces, reviewer roles and documented case handling for academic integrity teams.

Learn more →

Business

Protect brand, editorial and hiring workflows with SSO, audit logs and API automation.

Learn more →

Government

Residency options, retention controls and procurement support for public institutions.

Learn more →

Security & governance

Audit logs, configurable data retention, organization permissions and account security controls — because detection results end up in sensitive processes.

Review the security model →

API

Submit content, poll analysis status and retrieve structured evidence from your LMS, CMS or internal review tooling.

Read the API overview →

Pricing

Plans for individuals through government

Start free, upgrade when review volume grows, and move to Education, Business or Government terms when governance requirements arrive.

Free

$0/mo

Try the workflow with limited monthly scans.

Pro

Popular

$14.99/mo

For individuals who review content regularly.

Team

$49/mo

Shared workspace with reviewer workflow.

Education

$149/mo

Starting at — for schools and universities.

Business

$299/mo

Starting at — for workplaces and agencies.

Government

Contact

Procurement, residency and compliance support.

FAQ

Questions institutions ask first

Does TraceMark AI prove who wrote something?+

No. TraceMark AI estimates the likelihood that content shows characteristics associated with AI generation. It does not establish authorship, and every result requires human review.

What file types are supported?+

Pasted text, PDF, DOCX, TXT, JPG/JPEG and PNG.

Can results be wrong?+

Yes. All detection methods produce false positives and false negatives, particularly for short passages, translated text, and writing by non-native speakers.

How should institutions use results?+

As one input in a documented review process, alongside drafts, version history, interviews and subject-matter judgement.

Is my content used for training?+

No. Submitted content is processed for analysis only, with configurable retention and deletion.

How does the analysis work?+

Submitted text, documents and images are analysed by an AI model that reports likelihood, confidence, the signals it observed and the passages behind them. Results remain probabilistic estimates, never proof of authorship.

AI detection is probabilistic. This result does not establish authorship and should not be used as the sole basis for academic, employment, disciplinary, or legal decisions.