# Writing Style Checker (WSC)

> Free tool to detect weasel words, passive voice, duplicate words, long sentences, nominalizations, hedging, filler adverbs, and AI tells in technical writing.

## Overview

Writing Style Checker detects 8 types of writing issues:
- Weasel words (very, extremely, various, etc.)
- Passive voice (was written, has been done, etc.)
- Duplicate adjacent words (the the, is is, etc.)
- Long sentences (over 30 words by default)
- Nominalizations (implementation → implement, optimization → optimize)
- Hedging language (I think, it seems, etc.)
- Filler adverbs (totally, utterly, basically, etc.)
- AI tells (delve, tapestry, "let's dive in", etc.)

## Interfaces

- Web editor: https://wsc.theserverless.dev
- HTTP API: POST https://wsc.theserverless.dev/api/check
- Remote MCP server: https://wsc.theserverless.dev/mcp
- Local MCP server: npx wsc-mcp (npm package)
- CLI: npx wsc-cli check "**/*.md" (npm package)
- GitHub Action: theserverlessdev/wsc@master

## Documentation

- Full docs: https://wsc.theserverless.dev/docs
- Word library: https://wsc.theserverless.dev/words
- JSON Schema: https://wsc.theserverless.dev/schema.json
- Detailed LLM context: https://wsc.theserverless.dev/llms-full.txt

## API Quick Reference

POST /api/check
Content-Type: application/json

Request body:
- text (string, required): Text to analyze (max 100,000 chars)
- config (object, optional): Override detector settings

Response:
- summary: issue counts per detector and total
- issues: arrays of detected issues with word/phrase, index, line, column
- meta: characterCount, wordCount, processingTimeMs

## MCP Tools

- check_text: Analyze text for all 8 writing issues (accepts optional config)
- check_file: Read and analyze a file (local MCP only, auto-discovers .wscrc.json)
- fix_duplicates: Remove duplicate adjacent words
- list_word_lists: Return info about all detector word/phrase lists with counts

## Configuration

Create .wscrc.json in project root. Add "$schema": "https://wsc.theserverless.dev/schema.json" for editor autocompletion.

Detectors with word list overrides (add/remove): weaselWords, hedging, adverbs
Detectors with specialized config: longSentences (maxWords), nominalizations (add with word+suggestion)
AI tells detector: aiTells (add/remove for vocabulary, addPhrases/removePhrases for phrases)
Pattern-based detectors (enabled only): passiveVoice, duplicateWords

## Source

GitHub: https://github.com/theserverlessdev/wsc
License: MIT