# Unslop - **Website:** - **Primary alias:** `unslop.run` ## Description Unslop is an independent research project focused on identifying and improving machine-written text, particularly in scientific and academic writing. Its tools are designed to help users spot passages that may read as formulaic and revise them while preserving their original meaning. The site offers a text detector where users can paste or upload writing to see which sections appear machine-generated, as well as a rewriting tool that aims to retain citations, numbers, and technical terms. Users can also scan a webpage or an arXiv paper and review its content section by section. Alongside these tools, Unslop publishes notes about experiments in AI-text detection and large language models. The project is run by an individual researcher who trains models, conducts benchmarks, and turns the findings into publicly available tools. Unslop’s central idea is to expose or remove formulaic language without losing the substance of the writing, offering a practical resource for people preparing research papers and other technical material. ## Industries - 🖥 Computers Electronics and Technology - 🖥 Artificial Intelligence and Machine Learning _(under Computers Electronics and Technology)_ - 🧪 Science and Education ## Logos & icons - logo _(primary)_ — PNG — [download](https://cdn.brandfetch.io/idFYyvgM-A/w/180/h/180/theme/dark/logo.png?c=1bxid64Mup7aczewSAYMX&t=1790405438789) ## Colors | Name | Hex | Theme | | --- | --- | --- | | Copper | #c27c3a | accent | | Azure | #2b66ac | dark | | White | #ffffff | light | --- _Source: _ _See [/llms.txt](/llms.txt) for a full list of agent-readable pages, and [/auth.md](/auth.md) for how agents authenticate._