easywrite

Solid

Make AI-assisted academic and research writing read as human-authored while staying evidence-true, in English and 中文. Use for papers, theses, abstracts, related work, rebuttals/responses to reviewers, grant proposals (NSF/NIH/NSFC/面上/青年), cover letters, and any scholarly prose. Removes AI tells (em-dashes, "delve/underscore/tapestry", significance hype, rule-of-three, 综述腔/翻译腔/工程师腔), enforces claim↔evidence discipline (no verb stronger than its data), raises burstiness/perplexity so drafts stop tripping GPTZero/Turnitin/Originality.ai/CrossCheck, and matches the author's voice and the venue's register. Triggers: "去AI味", "说人话", "润色论文", "降AIGC率", "humanize", "lower the AI score", "make this less AI", "polish my paper", "reviewer response", "润色/改写/审稿 学术文本". NEVER alters a number, result, or citation, and NEVER fabricates sources; it is not a tool for evading AI-use disclosure.

AI & Automation 1 stars 0 forks Updated 1 weeks ago MIT

Install

View on GitHub

Quality Score: 78/100

Stars 20%
10
Recency 20%
90
Frontmatter 20%
70
Documentation 15%
100
Issue Health 10%
80
License 10%
100
Description 5%
100

Skill Content

# Easywrite Turn AI-assisted scholarly writing back into *a specific researcher making a precise, evidence-anchored argument* — not *a model performing "good academic writing."* Easywrite unifies five things a single humanizer usually misses: (1) the **claim↔evidence discipline** scholarship requires, (2) the **detector science** (perplexity / burstiness / stylometry) that decides whether clean prose gets falsely flagged, (3) a **bilingual EN + 中文** AI-tell catalog, (4) **voice and venue** matching, and (5) an **integrity firewall** so nothing becomes less true in the name of sounding human. > The whole thesis in one line: writing that is **human in voice** and **true in substance** — > easy to read, and easy to trust. --- ## IRON RULES (read first, never break) These override every stylistic instruction below. If a "humanizing" edit would violate one, don't make it. 1. **Never alter a number, statistic, result, equation, unit, or citation.** Not to round, not to "clean up," not to strengthen. Numbers and cites are load-bearing evidence, copy them verbatim. 2. **Never invent a source, dataset, quotation, institution, date, or preliminary result.** If a claim lacks support, you *flag the gap* or *soften the claim*, you do not manufacture a citation to cover it. 3. **No verb may be stronger than its evidence.** Empirical work *shows / provides evidence / is consistent with*; it does not *prove / demonstrate / guarantee* universal truths (Layer: [academic-disc...

Details

Author
Florent2025
Repository
Florent2025/easywrite
Created
1 weeks ago
Last Updated
1 weeks ago
Language
Python
License
MIT

Similar Skills

Semantically similar based on skill content — not just same category