yuwen-publish-precheck is a pre-publish review tool that uses AI for surface-level checks and contextual analysis based on official platform rules. It helps users determine if content can be published, identifies problematic sentences, and suggests edits. The tool supports platforms like Douyin, Xiaohongshu, and WeChat Channels, aiming to reduce compliance risks.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
An open-source tool called yuwen-publish-precheck performs pre-publishing risk review for Douyin, Xiaohongshu, and WeChat Channels using two-layer detection—lexical pre-check and AI context assessment—based on official platform rules, and provides modification suggestions.
yuwen-publish-precheck is a pre-publishing review tool. Users can have the AI review content before publishing; the tool tells whether the content can be published, which sentences are problematic, and how to modify them, based on official platform rules. It is applicable to domestic self-media platforms such as Douyin, Xiaohongshu, and WeChat Channels.
The tool employs a two-layer detection mechanism: lexical pre-check and AI context assessment, which is not simple banned-word matching. This means it can not only identify sensitive words but also assess risk based on context.
The tool's review rules come from 42 official rules and 72 original citations, ensuring judgments are evidence-based. Additionally, the judgment criteria are calibrated using 38 real approved manuscripts to improve accuracy.
The tool also has a memory function that remembers users' personal rules and safe expression libraries, enabling personalized risk avoidance.
The project is open-sourced on GitHub at github.com/yuwen-cool/yuwen-publish-precheck. Users can deploy it themselves or use it directly.
This report is based on a tweet from X user Geek, which is a product launch announcement and has not been independently verified. The tool's specific performance, accuracy, and other metrics have not been evaluated by third parties.
yuwen-publish-precheck provides self-media creators with an AI pre-review tool based on official rules, using two-layer detection and real-sample calibration, aiming to reduce content compliance risks.
Primary report
Primary source