Changelog

All notable changes to this project are documented in this file.

The format is based on Keep a Changelog, and
this project follows Semantic Versioning.

[2.1.0] - 2026-04-11

Added

• Corpus-trained humanization engine built from 10,000 Q1 open-access academic
  papers (OpenAlex, 11 domains, 2018–2025).
• Style model capturing real human writing patterns: sentence length distribution
  (mean 20.5, median 20, stddev 9.4), burstiness (0.426), vocabulary diversity
  (69.4%), passive voice ratio (18.1%), transition word frequencies, and sentence
  starter distributions.
• Style-aware prompt builder that injects corpus-calibrated statistical targets
  directly into LLM rewrite prompts.
• Corpus-calibrated detection thresholds — dynamic baselines derived from real
  academic writing rather than hardcoded values.
• Corpus-aware post-processing engine for output refinement.
• Expanded collocation database with 50+ new AI phrase replacements (150+ total).
• Bulk paper download and ingestion pipeline (scripts/papers/).

Changed

• Style model (public/corpus-style-model.json) now loads via fetch() for
  client-side compatibility (no node:fs in browser bundle).
• AI detector uses dynamic corpus-calibrated thresholds instead of static
  heuristics.

Fixed

• LinkedIn URL corrected in Footer.tsx.
• CI/CD workflow branch triggers updated from main to master.
• Client bundle crash resolved by isolating node:fs usage from browser-targeted
  style-model.ts.
• Markdown lint configuration fixed for Quality workflow.

[2.0.0] - 2026-03-31

Added

• 13 AI provider support with configurable API keys
• 4 rewrite levels including multi-pass ninja mode
• 13 preset writing tones and style controls
• Integrated AI detection heuristic scoring
• Readability analysis layer (Flesch-Kincaid, Gunning Fog)
• PDF and DOCX file upload support
• Grammar check integration
• Multi-language humanization support
• Browser-first API key storage (no server-side persistence)
• Side-by-side humanizer/detector workflow
• Dark/light theme toggle
• Research pipeline scripts for benchmark and training
• Comprehensive docs site with GitHub Pages deployment

Changed

• Major rewrite from initial prototype to production Next.js application
• Upgraded to Next.js 16, React 19, TypeScript 6
• Modernized CI with Node 20/22 matrix testing
• Overhauled UI with Tailwind CSS and glass-morphism design

Security

• All user data (API keys, history) stored client-side only
• No server-side persistence of user prompts or outputs

[1.0.0] - 2026-04-09

Added

• Comprehensive governance and support documentation (SUPPORT.md, improved
  contribution/security policies).
• Professionalized GitHub community templates and automation (CODEOWNERS,
  Dependabot, workflow modernization).
• Structured Markdown docs information architecture under docs/ for onboarding
  and usage.
• Release workflow for tag-based GitHub release publishing.

Changed

• Overhauled README.md for professional onboarding, architecture overview, and
  policy links.
• Modernized CI and Pages workflows for main branch and clearer quality gates.
• Added repository hygiene baselines (.editorconfig, .gitattributes, refined
  .gitignore).

Removed

• Removed duplicate root-level FUNDING.yml in favor of .github/FUNDING.yml.

What's Changed

• feat: add Chinese (Simplified & Traditional) language support for humanization by @rudra496 in #12
• Feat/chinese language support by @rudra496 in #13
• Copilot/develop advanced text analysis model by @rudra496 in #14
• PR4: Make training writes deterministic and preserve full experiment lineage by @Copilot in #15
• PR1 slice: file-level roadmap gap audit for PR1–PR9 by @Copilot in #16
• PR8 hardening: make CI lint/build deterministic and enforce PR2 integrity/schema gates by @Copilot in #18
• chore: safe full repository modernization for docs, governance, CI/CD, and v1.0.0 readiness by @Copilot in #17
• build(deps-dev): bump tailwindcss from 3.4.19 to 4.2.2 by @dependabot[bot] in #19
• build(deps): bump lucide-react from 0.468.0 to 1.8.0 by @dependabot[bot] in #26
• build(deps): bump react and @types/react by @dependabot[bot] in #27
• build(deps-dev): bump eslint-config-next from 14.2.20 to 16.2.3 by @dependabot[bot] in #28
• build(deps-dev): bump postcss from 8.5.8 to 8.5.9 by @dependabot[bot] in #23
• build(deps-dev): bump typescript from 5.9.3 to 6.0.2 by @dependabot[bot] in #21
• build(deps-dev): bump @types/node from 22.19.17 to 25.5.2 by @dependabot[bot] in #25
• build(deps): bump react-dom and @types/react-dom by @dependabot[bot] in #22
• build(deps-dev): bump eslint from 8.57.1 to 10.2.0 by @dependabot[bot] in #24
• build(deps): bump next from 14.2.20 to 16.2.3 by @dependabot[bot] in #20
• fix: resolve CI/Vercel production failures from unmitigated major dependency upgrades by @Copilot in #29
• Fix integration smoke by bootstrapping missing benchmark/model artifacts by @rudra496 in #30
• Add batch paper ingest+train pipeline, fallback seeding, configs, and smoke test by @rudra496 in #31
• Add one-command Q1-ready pipeline script and npm shortcuts; document in README by @rudra496 in #39
• build(deps): bump actions/setup-node from 4 to 6 by @dependabot[bot] in #32
• build(deps): bump actions/deploy-pages from 4 to 5 by @dependabot[bot] in #33
• build(deps): bump actions/configure-pages from 5 to 6 by @dependabot[bot] in #34
• build(deps): bump actions/checkout from 4 to 6 by @dependabot[bot] in #35
• build(deps): bump actions/upload-pages-artifact from 3 to 4 by @dependabot[bot] in #36
• build(deps): bump DavidAnson/markdownlint-cli2-action from 17 to 23 by @dependabot[bot] in #37

New Contributors

• @rudra496 made their first contribution in #12
• @Copilot made their first contribution in #15
• @dependabot[bot] made their first contribution in #19

Full Changelog: v2.0.0...v2.1.0

🥷 StealthHumanizer v2.0.0 — Major Release

What's New

• 13 AI Providers (10 free): Gemini, Groq, ZAI, OpenRouter, Together, Cerebras, Mistral, Cohere, DeepInfra, HuggingFace, Cloudflare, OpenAI, Claude
• 4 Rewrite Levels: Light, Medium, Aggressive, Ninja (5-pass auto-loop)
• 13 Tone Presets: Conversational, Academic, Journalistic, Creative, Professional, Technical, Persuasive, Storytelling, Humorous, Emotional, Analytical
• 5 Writing Styles: Academic, Professional, Casual, Creative, Technical
• 12-Metric AI Detector with readability scoring
• Multi-Pass Humanization: Auto re-humanizes until target score reached
• GitHub Pages: Standalone version at https://rudra496.github.io/StealthHumanizer/
• Full SEO: Open Graph, Twitter Cards, JSON-LD, sitemap, robots.txt
• GitHub Actions: CI build check + auto Pages deploy
• Security: SECURITY.md, FUNDING.yml, privacy-first architecture

Free API Keys to Get Started

• Google Gemini — Best quality, free
• Groq — Ultra-fast, free
• ZAI GLM-5 — Powerful, free

Installation

git clone https://github.com/rudra496/StealthHumanizer.git
cd StealthHumanizer && npm install && npm run dev

Or use online: https://rudra496.github.io/StealthHumanizer/

🔒 Privacy

All API keys stored only in your browser. No server, no tracking, no accounts.

Built by Rudra Sarker | MIT License