alex

Cloud-hosted writing assistants require transmitting repository documentation to external servers and rarely integrate cleanly into build automation. alex functions as a local command-line utility and Node.js library that inspects Markdown, HTML, and MDX files directly inside your development workflow. It catches gender-favoring, ableist, and polarizing phrasing at the pull request stage without exposing private text over the network.

Where alex fits

While consumer text tools operate through browser extensions or closed desktop apps, technical documentation requires automated validation within continuous integration pipelines. General prose linters often demand writing complex regular expressions or maintaining custom YAML style rules just to flag common non-inclusive terms. alex delivers an out-of-the-box ruleset powered by the unified and retext syntax tree ecosystem. By parsing Markdown and MDX syntax trees rather than treating files as raw strings, it inspects text nodes while ignoring inline code spans, code blocks, and URLs where technical terms would trigger false positives.

The tool focuses strictly on equality and considerate language rather than full-spectrum proofreading. It will not flag spelling mistakes, grammatical errors, or awkward phrasing. Instead of attempting to replace a human copy editor, it serves as a lightweight gatekeeper that developers run via npm scripts, GitHub Actions, or pre-commit hooks. Engineering teams typically pair alex with dedicated spell-checkers, using its CLI flags to stop exclusionary terms from reaching public documentation.

What it does well

• Syntax-aware linting — parses Markdown, MDX, and HTML into abstract syntax trees, checking prose while skipping code blocks, inline code, and URLs.
• Git diff inspection — supports a --diff flag that checks only modified lines in pull requests, preventing legacy warnings from breaking active builds.
• Inline comment control — lets writers silence specific rules or disable checks across entire sections using comments like ``.
• Flexible configuration — loads settings from .alexrc files in JSON or YAML, JavaScript configs, or package.json, supporting granular allow and deny lists.
• Tunable profanity detection — incorporates the cuss dictionary with configurable sureness levels from zero to two to distinguish probable profanities from common words.