# theserverlessdev/wsc [Health: Active]

**Category:** 📊 Monitoring  
**Repository:** https://github.com/theserverlessdev/wsc  
**GitHub Stars:** 9  
**npm Downloads (last month):** 36  
**Views:** 4  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/theserverlessdev-wsc

## Description
Prose linter + AI-slop detector: weasel words, passive voice, hedging, and 190+ research-cited AI tells. Remote server at wsc.theserverless.dev/mcp or npx wsc-mcp.

## Tools
Capabilities this server exposes over MCP:

- **check_text** — Analyze text for all 8 writing style issues. Accepts optional `config`.
- **fix_duplicates** — Remove duplicate adjacent words and return cleaned text
- **list_word_lists** — Return info about all detector word lists
- **check_file** — *(Local only)* Read a file from disk and analyze it. Auto-discovers `.wscrc.json`.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "wsc": {
    "command": "npx",
    "args": ["-y","wsc-mcp"]
  }
}
```

## Documentation

## What theserverlessdev/wsc MCP server does

The theserverlessdev/wsc MCP server gives MCP clients access to a prose analysis engine focused on specific, explainable writing patterns. It checks text for eight detector categories: weasel words, passive voice, adjacent duplicate words, long sentences, nominalizations, hedging, filler adverbs, and AI-associated words, phrases, and sentence structures.

Results identify detected issues rather than claiming to prove who wrote the text. The AI-tell lists are based on published corpus research and other cited sources, while the traditional style lists cover common editing concerns such as vague wording and unnecessarily weak assertions.

## How it works

A remote endpoint is available at `https://wsc.theserverless.dev/mcp` and uses Streamable HTTP, so a compatible MCP client can connect without installing the package. The local option runs through npm with `npx wsc-mcp` and communicates over stdio.

The main `check_text` tool accepts text and an optional configuration object, then returns structured analysis for all eight detectors. `fix_duplicates` handles one narrow correction by removing repeated adjacent words without requiring a full analysis workflow. `list_word_lists` reports information about the detector lists. `check_file` is available only in the local server and reads a file from disk before analyzing it.

## Setup and configuration

For a remote connection, add the MCP URL to the client’s server configuration. For local use, run `npx wsc-mcp`; this is also the configuration pattern documented for Claude Desktop and Claude Code. The local package is useful when an agent needs to inspect files on the same machine.

The theserverlessdev/wsc MCP server supports optional detector settings through a `.wscrc.json` file or a configuration object supplied to analysis. Settings can enable or disable detectors, add or remove entries from supported lists, and change the maximum word count for long-sentence detection. Missing settings use defaults. The local file tool automatically discovers `.wscrc.json`.

## Tools and capabilities

- `fix_duplicates`: remove adjacent repeated words and return the cleaned text.
- `list_word_lists`: return metadata about the detector word lists.
- `check_file`: read and analyze a local file; available only through the local server.

The detector inventory includes 95 weasel-word entries, 260 irregular verbs for passive-voice checks, 245 nominalization pairs, 100 hedging phrases, 139 filler adverbs, and AI-tell entries covering words, inflected forms, phrases, and structural patterns. Long sentences use a 30-word default threshold, and duplicate matching ignores case and whitespace differences.

## Limitations and notes

Remote and MCP processing handles only text explicitly sent to the endpoint. The local-only `check_file` capability is not available through the hosted server. The service is designed to flag concrete patterns and does not establish authorship; AI-writing indicators can occur in human-written prose as well. Configuration is optional, and detector behavior can change when custom entries or thresholds are supplied.

_Full upstream README: https://allmcps.com/mcp/theserverlessdev-wsc/readme_