Summary
SlopTotal detects AI-generated text using 23 local analysis engines and produces a calibrated score.
About
SlopTotal provides locally operated AI-text detection for organizations that want to assess text without sending it to external APIs. Its multi-engine approach may support review workflows, but detection scores should be treated as indicators rather than definitive proof of authorship.
SlopTotal is an open-source AI text detector that runs 23 independent engines in parallel on a local CPU. It combines neural classifiers, perplexity measures, and heuristics to produce a calibrated detection score without API calls. The tool is designed to run with 4–8 GB of RAM and includes a Docker setup and live demo. Its detection approach has been tested on prose written before 1920 to address false positives.
Key Features
Runs 23 independent detection engines in parallel
Combines neural classifiers, perplexity measures, and heuristics
Produces a calibrated AI-text detection score
Runs locally on CPU with 4–8 GB of RAM and makes no API calls
Includes a Docker setup and has been tested on pre-1920 prose to address false positives
Real-World Use Cases
Editorial review
Publishers and editorial teams can use detection scores to flag submissions for additional review, while retaining human judgment for final decisions.
Education and academic integrity
Educators can use the tool as one input when reviewing student work, with care to avoid treating detector results as conclusive evidence.
Privacy-conscious content screening
Organizations can assess text locally without transmitting it to an external detection API.
Detection research and evaluation
Researchers and developers can examine a multi-engine detection approach and test it against varied text, including historical prose.
Business Opportunities
Private content-review workflows
Problem: Organizations may need to screen text for possible AI generation while limiting disclosure of unpublished or sensitive material to external services.
Solution: Integrate the self-hosted detector into an internal review process, using its score to prioritize human assessment rather than make automatic authorship decisions.
Estimated ROI: Potentially moderate through reduced manual triage and lower reliance on external APIs; actual returns depend on detection accuracy, review volume, and integration effort.
Execution: Deploy the tool locally using its Docker setup, define review thresholds through internal testing, and document human-review and appeal procedures.
AI-detection evaluation services
Problem: Teams adopting AI-text detectors need to understand how detection results behave across content types and how often false positives occur.
Solution: Offer evaluation and implementation support using local tests, including checks on human-written and historical text, and advise clients on appropriate use of scores.
Estimated ROI: Potentially moderate for specialist consulting, contingent on client demand, evaluation rigor, and the ability to demonstrate practical value.
Execution: Build representative test sets with client consent, compare detector outputs against known text origins, and provide deployment and policy guidance without presenting scores as proof.