Sapling MCP Server

Use Sapling's grammar checking, AI detection, paraphrasing, summarization, sentiment analysis, language detection, and zero-shot text classification tools from Claude and other MCP-compatible assistants.

Setup

Create a Sapling API key from https://sapling.ai/docs/api/api-access. Node.js 18 or newer is required.

Claude Desktop

Add this to your Claude Desktop config:

{
  "mcpServers": {
    "sapling": {
      "command": "npx",
      "args": ["-y", "@saplingai/mcp-server"],
      "env": {
        "SAPLING_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Claude Code

claude mcp add sapling --env SAPLING_API_KEY=YOUR_API_KEY \
  -- npx -y @saplingai/mcp-server

Environment

• SAPLING_API_KEY: Required Sapling API key.
• SAPLING_API_BASE_URL: Optional API base URL. Defaults to https://api.sapling.ai.
• SAPLING_MCP_TIMEOUT_MS: Optional request timeout. Defaults to 120000.

Tools

• sapling_grammar_check: Grammar, spelling, punctuation, and style suggestions.
• sapling_spell_check: Spelling-focused suggestions.
• sapling_ai_detect: AI-generated text detection.
• sapling_rephrase: Paraphrase or transform text.
• sapling_summarize: Summarize text.
• sapling_extract: Structured data extraction — name 1-20 fields (plain names or {name, type, description, required} objects, with type one of string, number, integer, boolean, date or list) and get back a data map of typed values, a per-field breakdown with the verbatim evidence span each value came from, and the names the text didn't yield in missing. Optional context describes what the document is. Values must be grounded in the text, so nothing is invented.
• sapling_sentiment: Sentiment analysis.
• sapling_language_detect: Language detection — top language with name and confidence, runner-up candidates (top_k), and an optional per-segment breakdown for mixed-language text (segments).
• sapling_classify: Zero-shot text classification — assign one (or several, with multi_label) of your own labels (plain names or {name, description} objects, 2–20) to a text; returns the best label, applicable labels, per-label scores and a one-sentence rationale. Optional context describes what the texts are.
• sapling_seo: SEO analysis and metadata — readability stats, target-keyword density (keywords), the page's top terms, and LLM-generated title tags, meta descriptions, slug and focus keywords; suggestions: false gives the free stats-only mode.
• sapling_quality: Writing quality score (1–5), optionally per sentence (sentence_scores) and with an LLM-judged rubric — clarity/coherence/correctness/concision, summary, quoted issues with suggested rewrites (rubric, billed at a higher rate).
• sapling_detect_pii: PII detection and redaction — emails, phone numbers, US SSNs, credit cards, IP addresses, IBANs and US bank routing numbers with exact character offsets; redact: true also returns redacted_text with [TYPE] placeholders.
• sapling_safety: Content-safety scores — toxicity, profanity, harassment, hate speech, self-harm, sexual content and violence, each scored 0–1 with flagged categories (threshold, default 0.5).

Development

npm ci
npm test