Add humanize-chinese to Demos & Products (Chinese AIGC detector)

[voidborne-d]

Follow-up to issue #5. Adding humanize-chinese to Detection → Demos & Products.

What it is

An open-source Python library for detecting AI-generated Chinese text and rewriting it to reduce AIGC signals. Pure Python, zero external dependencies — runs locally, no API calls, no GPU required.

Why it fits here

The current Demos & Products list is strong on English detectors (GPTZero, Sapling, Writefull, etc.) and multi-lingual commercial products, but has no open-source Chinese-specific detector. humanize-chinese fills that gap:

• Chinese-specific statistical features — 48 patterns that differ from English AI-writing fingerprints: 四字成语 density, 并列对称 preference, 的地得 over-correctness, connective-word front-loading, abstract-noun inflation. These are signals that perplexity-based English detectors miss when tokenizing Chinese text.
• Open source — MIT, all code inspectable, unlike the closed-source commercial detectors that dominate the current list.
• Academic AIGC workflow support — CNKI/VIP/Wanfang-oriented scoring and reduction heuristics for the 论文降重 use case.
• Interpretable output — 0-100 score with per-feature breakdown at sentence level, not a black-box classifier.

Diff

One-line addition at the end of ### Demos & Products, matching the existing entry format.

Happy to adjust the wording, placement, or format if you'd prefer. Thanks for maintaining this list!