Score every page for readability, structure, completeness, and Diataxis classification. API or GitHub Action. Results in seconds.
curl -X POST https://docs-score.com/v1/docs/score-url \
-d '{"url": "https://your-docs.dev/getting-started"}'
Send a docs URL or raw markdown to the API. One endpoint, one POST request.
Get scores across 5 quality dimensions plus automatic Diataxis classification (Tutorial, Guide, Reference, Explanation).
Use scores in CI to catch regressions, or build dashboards to track docs health over time.
{
"url": "https://your-docs.dev/getting-started",
"overall_score": 0.74,
"dimensions": {
"completeness": 0.80,
"readability": 0.68,
"structure": 0.85,
"accuracy_signals": 0.60,
"diataxis": "tutorial"
},
"recommendations": [
"Flesch-Kincaid grade level 14.2 — aim for 8-10",
"3 sections lack code examples",
"2 broken links detected"
]
}Five dimensions, each scored 0.0 to 1.0. Composite weighted score determines pass/fail.
| Dimension | What it measures |
|---|---|
| Completeness | Coverage of expected sections, prerequisites, examples, and edge cases |
| Readability | Flesch-Kincaid grade level, sentence length, passive voice ratio |
| Structure | Heading hierarchy, code block presence, list usage, section balance |
| Accuracy signals | TODO markers, broken links, outdated version references, missing code examples |
| Diataxis | Automatic classification: Tutorial, How-to Guide, Reference, or Explanation |
Fail the build when docs quality drops below threshold. Catch regressions before they ship.
Track every page's score over time. See which sections are improving and which are degrading.
Score your entire docs site before and after a migration. Prove quality was preserved.
Give writers objective feedback. Replace subjective review cycles with measurable quality scores.
Add docs quality checks to any PR. Scores run on changed markdown files.
name: Docs Quality
on: [pull_request]
jobs:
score:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Score docs
run: |
for f in $(git diff --name-only HEAD~1 -- '*.md'); do
curl -s -X POST https://docs-score.com/v1/docs/score \
-H "Authorization: Bearer ${{ secrets.DOCS_SCORE_KEY }}" \
-H "Content-Type: application/json" \
-d "{\"content\": \"$(cat $f | jq -Rsa .)\"}" \
| jq '.overall_score'
donedocs-score is the Docs vertical of Proofmark — the quality gate for AI responses. Same scoring engine, purpose-built for documentation.