Back to Tool
S

SlopTotal

Tool Nov 3, 2026

A self-hosted AI-text detection application combining 23 engines with calibrated scores, document scanning and inspectable per-engine results.

Visit primary link 2 sources attached
SlopTotal author demo shows an empty text input, document and webpage tabs, a sample button and an Analyze text control.
First frame from the project-supplied demonstration in the English v1.3.0 README. It illustrates the detector interface, not an AITrending test or proof of detection accuracy. Source: pablocaeg / SlopTotal contributors
self-hosted

What happened lately

SlopTotal timeline

SlopTotal runs neural classifiers, statistical tests and linguistic heuristics on text, URLs and supported documents. It is a standalone application with a web interface and API, with an associated Chrome extension. Version 1.3.0 adds CUDA inference and live score bounds.

The application is MIT-licensed; bundled model weights keep their own terms. Docker and Python source installation are documented. Self-hosted inference requires no third-party AI API account, but model downloads and sufficient local memory are required. A scan submitted to a public instance runs on that instance’s server.

The project’s own evaluation identifies false positives, weaker results on non-native English essays, and unreliable short-text signals. Its scores are indicators to inspect, not proof that a particular person used AI.