Score text readability (Flesch, SMOG, ARI, FK Grade) via mcp-readability (npx)

Recipe: Score text readability via mcp-readability (npx)

Server

npx -y mcp-readability — stdio, NDJSON framing, Node.js, zero config, no API key.

Tools (1)

Parameters

• text (string, required) — the text to analyze. Minimum ~2 sentences for reliable scores.
• include_suggestions (boolean, optional, default true) — when true, returns suggestions for long sentences (>25 words), complex words (>4 syllables), and passive voice.

Example: Simple text (middle-school level)

Request:

{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"analyze_readability","arguments":{"text":"The quick brown fox jumps over the lazy dog. Reading level analysis helps writers ensure their content is accessible to their target audience."}}}

Response:

{
  "scores": {"flesch_reading_ease": 55.4, "flesch_kincaid_grade": 8.4, "smog_index": 8.5, "ari": 8.5},
  "reading_level": "Middle School (6th–8th grade)",
  "audience": "Middle school students",
  "stats": {"word_count": 23, "sentence_count": 2, "syllable_count": 38, "avg_words_per_sentence": 11.5, "avg_syllables_per_word": 1.65},
  "summary": "Flesch Reading Ease: 55.4 | FK Grade: 8.4 | SMOG: 8.5 | ARI: 8.5 — Middle School (6th–8th grade) (23 words, 2 sentences)",
  "suggestions": ["No major readability issues found — text looks good for the target audience."]
}

Example: Dense academic text (graduate level)

Request:

{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"analyze_readability","arguments":{"text":"The implementation of machine learning algorithms in production environments necessitates consideration of computational complexity, algorithmic efficiency, and the inherent trade-offs between model accuracy and inference latency. Furthermore, the deployment pipeline must accommodate continuous integration and continuous deployment paradigms while maintaining rigorous statistical validation of model performance metrics across heterogeneous data distributions.","include_suggestions":true}}}

Response:

{
  "scores": {"flesch_reading_ease": -52.9, "flesch_kincaid_grade": 27, "smog_index": 24.2, "ari": 28.9},
  "reading_level": "Graduate (17th grade+)",
  "audience": "Graduate students / specialists",
  "stats": {"word_count": 51, "sentence_count": 2, "syllable_count": 141, "avg_words_per_sentence": 25.5, "avg_syllables_per_word": 2.76},
  "suggestions": [
    "1 sentence exceeds 25 words — consider splitting.",
    "5 words have 5+ syllables — consider simpler alternatives: implementation, necessitates, consideration, computational, heterogeneous"
  ]
}

Gotchas

• Needs at least 2 sentences for reliable scores. Single sentences produce noisy results.
• Negative Flesch Reading Ease is valid — means extremely difficult text.
• Server uses NDJSON framing (newline-delimited), not Content-Length HTTP framing.
• Pure in-process computation — no network calls, no API key, works offline.