Humanizer app for Sero — transform text with creative writing styles and remove AI-generated patterns. Includes a comprehensive Pi skill based on Wikipedia's "Signs of AI writing" guide.
Install in Sero → Admin → Plugins with:
git:https://github.com/monobyte/sero-humanizer-plugin.git
Sero clones the source repo, installs its dependencies locally, builds the UI, and then hot-loads the plugin into the sidebar.
Install as a Pi package:
pi install git:https://github.com/monobyte/sero-humanizer-plugin.gitThe agent gains a humanize tool and a /humanize command, plus the
humanizer skill — a detailed writing-editor prompt that detects and fixes
AI writing patterns including inflated symbolism, promotional language, vague
attributions, em-dash overuse, rule of three, AI vocabulary, and more.
- AI Pattern Detection — identifies common AI writing tells
- Style Transformation — rewrites text to sound natural and human-written
- History Tracking — saves transformation history for review
- Pi Skill — comprehensive skill definition (
skills/humanizer/SKILL.md) that teaches the agent to apply Wikipedia's AI-cleanup guidelines - Web UI — visual editor with side-by-side original/transformed panes, style presets, and history panel
When loaded in Sero, the web UI mounts in the main app area. Paste or type text, choose a writing style, and the agent transforms it using the humanizer skill. Transformation history is persisted and visible in both the UI and via the agent tool.
~/.sero-ui/
└── apps/
└── humanizer/
└── state.json
{
"history": [
{
"id": "abc123",
"original": "The text delves into...",
"humanized": "The text examines...",
"style": "casual",
"timestamp": "2026-01-01T00:00:00.000Z"
}
]
}The plugin ships a Pi skill at skills/humanizer/SKILL.md (v2.2.0) that
covers:
- Inflated symbolism & promotion detection
- Superficial
-inganalyses - Vague attributions and weasel words
- Em-dash and semicolon overuse
- Rule-of-three and list padding
- AI vocabulary replacement (delve, tapestry, landscape, etc.)
- Negative parallelisms and excessive conjunctive phrases
The skill is auto-registered via pi.skills: ["./skills"] in package.json.
npm install
npm run dev # Vite dev server (port 5192)
npm run build # Production build → dist/ui/
npm run typecheck # TypeScript check