Axis A: Readability
Ten standard formulas (Flesch Reading Ease, Flesch-Kincaid, Gunning Fog, SMOG, Coleman-Liau, and more) plus a consensus grade level. Answers: is this pitched at the right reading level for its audience?
docstats
Score readability and lint technical prose for house style. Run it as an MCP tool for AI coding assistants, or as a REST service in your pipeline.
Two Independent Axes
Section titled “Two Independent Axes”docstats evaluates documents along two independent dimensions and reports both without blending them into a single score.
Axis B: House-Style Linting
Deterministic pattern checks for throat-clearing openers, binary contrast frames, non-technical filler adverbs, and rhetorical em dashes. Answers: is the prose crisp and direct?
Post-Hoc Acceptance Gating
Section titled “Post-Hoc Acceptance Gating”docstats operates as an asynchronous acceptance gate for CI/CD pipelines, PR reviews, and pre-publish QA. In empirical benchmarks, injecting live numeric metrics during text generation yielded no measurable quality improvement over textual guidance (p = 0.7253) and induced metric gaming. Authors draft against clear guidelines and audit completed text with docstats.
Documentation
Section titled “Documentation”Getting StartedInstall with uv and run your first analysis in under a minute.
MCP & Agent PluginsWire docstats into Claude Code, Gemini CLI, Cursor, and other agent runtimes.
Interpreting ScoresRead the two-axis scorecard and turn numbers into edits.
Readability FormulasA deep dive into the linguistics behind each of the ten metrics.