Create-knowledge-find-decode/PRODUCT.md

Product

Register

product

Users

Content knowledge builders who need to inspect, classify, and structure creator-facing knowledge from Xiaohongshu, WeChat, and Douyin results. They work in a dense review workflow where source media, model judgments, and extraction state must be visible together.

Product Purpose

This tool turns searched posts, articles, and videos into inspectable creation-knowledge candidates, then helps operators review whether each item teaches transferable content-creation decisions. Success means users can quickly see query coverage, platform health, media availability, and AI judgment outcomes without losing the original evidence.

Brand Personality

Precise, restrained, operational.

Anti-references

Avoid marketing-style landing pages, decorative dashboards, oversized hero sections, and vague AI-generated explanations. The interface should not hide evidence behind ornamental cards or make status hard to scan.

Design Principles

• Evidence first: media, source text, and judgment reason stay close together.
• State is data: done, failed, partial, unjudged, creation, and non-creation states must be explicit.
• Dense but readable: favor compact layouts that support repeated review.
• Conservative judgment: ambiguous creation-knowledge boundaries should remain visible rather than silently discarded.

Accessibility & Inclusion

Use system fonts, clear semantic color plus text labels, and responsive layouts that keep media cards readable on narrow screens. Do not rely on color alone for classification state.