Aspirin0000/zhouli-translator

What it solves

This project provides a way to transform ordinary modern Chinese into a stylized, pseudo-ancient "Zhouli" style (a formal, scholarly translation tone) and vice versa. It allows users to turn mundane daily events into formal, ritualistic arguments and decode these stylized texts back into plain, direct language.

How it works

The application uses the DeepSeek LLM (specifically deepseek-v4-flash) to perform these transformations. It employs specific prompt construction and cleaning rules to handle two directions:

• Ask Ritual (问礼): Rewrites modern text into a formal, scholarly tone that frames modern issues within the context of ritual, status, and propriety.
• Explain Ritual (释礼): Strips away the formal packaging to reveal the original, direct meaning of the text.

The system supports multiple modes (e.g., "Great Scholar" or "Sharp Critique") and three length settings (Short, Standard, and Long) to control the output style.

Who it’s for

• Users who want to create humorous, meme-like content in a specific ancient-style Chinese writing style.
• Developers looking for an example of a Next.js application integrated with LLM APIs and Cloudflare Workers/D1 for analytics.

Highlights

• Bidirectional Translation: Converts between plain Chinese and "Zhouli" style.
• Customizable Styles: Offers various perspectives for both generation and explanation.
• Length Control: Three distinct output length levels for both directions.
• Image Export: Allows users to save results as "Ritual Cards" or "Explanation Cards."
• Skill Package: Includes a standalone speak-zhouli Skill for easier distribution.
• Integrated Analytics: Uses Cloudflare D1 to track interaction metrics and feedback anonymously.