Grammarly MCP server for AI agents
Let Claude, ChatGPT or your own agent read and act on Grammarly: 17 typed functions through one MCP server. No Grammarly API key needed.
- Site
- grammarly.com
- Functions
- 17
- Sign-in
- For account actions
- Works with
- MCP, HTTP, Node, Python
Connect
An MCP server for Grammarly alone, with one tool per function. Paste it into Claude, ChatGPT, Cursor or any MCP client as a remote server.
https://api.bowmark.ai/mcp/site/grammarlyOne MCP server for Grammarly and every other site Bowmark covers. Your agent looks up what it needs and calls it.
https://api.bowmark.ai/mcpSet it up in your clientBoth need a Bowmark API key. The quickstart gets you one and a first result in a few minutes. Functions that need a login use the person's own account: how sign-in works.
Functions
Open a function to see its arguments, what it returns and how to call it.
checkGrammarChecks text for grammar, spelling and punctuation problems and returns each alert with its position and suggested replacements.Arguments
Name Type Required textstringThe text to check, up to 150,000 characters.Yes Returns
GrammarCheckinterface GrammarAlert { /** "Grammar", "Punctuation", "Spelling", "Clarity", … */ group: string; category: string; title: string; /** Character offsets into the submitted text, end exclusive. */ begin: number; end: number; highlightText: string; /** Suggested replacements; empty when Grammarly withholds the fix. */ replacements: string[]; explanation: string; impact: string | null; /** True for a Premium-only alert: the problem is named, the fix withheld. */ premium: boolean; } interface GrammarCheck { alerts: GrammarAlert[]; /** Alert counts per outcome: Correctness, Clarity, Engagement, Tone. */ outcomeAlerts: Record<string, number>; /** 0-100, or null when the text is too short to score. */ score: number | null; premiumAlertsCount: number; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/checkGrammarCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.checkGrammar({ text: "She go to school yesterday." });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.checkGrammar({"text": "She go to school yesterday."}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/checkGrammar \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"text":"She go to school yesterday."}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.checkGrammar({ text: "She go to school yesterday." });countWordsReturns word/character/sentence/paragraph counts, reading and speaking time, and a Flesch Reading Ease readability score.Arguments
Name Type Required textstringYes Returns
WordCountStatsinterface TimeEstimate { h: number; m: number; s: number; } interface WordCountStats { words: number; /** Character count including whitespace. */ characters: number; charactersNoSpaces: number; sentences: number; paragraphs: number; readingTime: TimeEstimate; speakingTime: TimeEstimate; /** Flesch Reading Ease, 0-100 (higher is easier); null for empty text. */ readabilityScore: number | null; readabilityDescription: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/countWordsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.countWords({ text: "Two roads diverged in a yellow wood, and sorry I could not travel both." });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.countWords({ "text": "Two roads diverged in a yellow wood, and sorry I could not travel both." }))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/countWords \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"text":"Two roads diverged in a yellow wood, and sorry I could not travel both."}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.countWords({ text: "Two roads diverged in a yellow wood, and sorry I could not travel both." });detectAIScores 0-100 how much of a text Grammarly's AI detector judges AI-written, with the passages behind the score.Arguments
Name Type Required textstringYes Returns
AIDetectioninterface AIDetection { /** 0-100: the share of the text judged AI-written. */ score: number; /** The passages behind the score, as character offsets (end exclusive). */ ranges: { begin: number; end: number; score: number }[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/detectAICall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.detectAI({ text: "<text>" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.detectAI({"text": "<text>"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/detectAI \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"text":"<text>"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.detectAI({ text: "<text>" });detectPlagiarismCounts how many passages of a text Grammarly's free scan matches against web and academic sources, alongside its other issue counts.Arguments
Name Type Required textstringYes Returns
PlagiarismCheckinterface IssueCount { /** "Plagiarism", "Spelling", "Style", … */ group: string; category: string; label: string; count: number; } interface PlagiarismCheck { /** Passages matched against web pages and academic sources; 0 = none found. */ plagiarismCount: number; /** Every count the free scan returns, counts only, not the matching sources. */ issues: IssueCount[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/detectPlagiarismCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.detectPlagiarism({ text: "<text>" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.detectPlagiarism({"text": "<text>"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/detectPlagiarism \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"text":"<text>"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.detectPlagiarism({ text: "<text>" });detectToneReads how a text sounds to a reader with Grammarly's tone detector, e.g. appreciative, optimistic or disapproving, each with a confidence.Arguments
Name Type Required textstringYes Returns
ToneDetectioninterface DetectedTone { /** Grammarly's tone label, e.g. "appreciative", "optimistic", "disapproving". */ name: string; emoji: string; /** 0-1, how sure the detector is. */ confidence: number; } interface ToneDetection { /** Strongest first. Empty when the text is too short or carries no clear tone. */ tones: DetectedTone[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/detectToneCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.detectTone({ text: "<text>" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.detectTone({"text": "<text>"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/detectTone \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"text":"<text>"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.detectTone({ text: "<text>" });getAccountReturns the signed-in account's profile: name, email, plan (Free or Premium) and whether the email is confirmed.Arguments
None.
Returns
GrammarlyAccountinterface GrammarlyAccount { /** The signed-in account's email. */ email: string; /** The account's display name, or "" when none is set. */ name: string; /** The site's own plan name, "Free" or "Premium". */ plan: string; confirmed: boolean; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/getAccountCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.getAccount();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.getAccount())curl -X POST https://api.bowmark.ai/v1/sites/grammarly/getAccount \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.getAccount();getDocumentReturns one saved document's title and full plain text from the caller's own account, by the id listDocuments returns.Arguments
Name Type Required idstringThe document id listDocuments returned.Yes Returns
GrammarlyDocumentTextinterface GrammarlyDocumentText { id: string; /** From the download's filename, or "" when the site sent none. */ title: string; /** The document's full plain text. */ text: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/getDocumentCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.getDocument({ id: "1" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.getDocument({"id": "1"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/getDocument \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"id":"1"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.getDocument({ id: "1" });humanizeRewrites AI-sounding text to read more naturally with Grammarly's free AI humanizer.Arguments
Name Type Required textstringYes Returns
Humanizedinterface Humanized { text: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/humanizeCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.humanize({ text: "<text>" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.humanize({"text": "<text>"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/humanize \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"text":"<text>"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.humanize({ text: "<text>" });listDocumentsLists the saved documents in the caller's own Grammarly account, with title, preview, issue count and dates.Arguments
Name Type Required searchstringOnly documents whose title or text contains this.No limitnumberAt most this many. Omit for all.No trashbooleanList the trash instead of the live documents.No Returns
GrammarlyDocumentsinterface GrammarlyDocument { id: string; title: string; /** The opening of the document's text. */ preview: string; /** Open issues the checker counted. */ errors: number; size: number; /** The uploaded file's name, or null for a document typed in the editor. */ originalFilename: string | null; createdAt: string | null; updatedAt: string | null; } interface GrammarlyDocuments { documents: GrammarlyDocument[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/listDocumentsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.listDocuments({});import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.listDocuments({}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/listDocuments \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.listDocuments({});listGoalsLists the writing goals (goals, domain, audience, formality, dialect) set on each saved document in the caller's own account.Arguments
Name Type Required documentIdstringOnly this document's goals (an id listDocuments returned).No Returns
GrammarlyGoalsinterface GrammarlyDocumentGoals { documentId: string; title: string; /** Subset of "inform" · "describe" · "convince" · "tell_a_story" · "sell"; [] when none. */ goals: string[]; /** "academic" · "business" · "general" · "mail" · "casual" · "creative", or null. */ domain: string | null; /** "general" · "knowledgeable" · "expert", or null. */ audience: string | null; /** "informal" · "neutral" · "formal", or null. */ formality: string | null; /** "american" · "british" · "canadian" · "australian" · "indian", or null. */ dialect: string | null; } interface GrammarlyGoals { documents: GrammarlyDocumentGoals[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/listGoalsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.listGoals({});import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.listGoals({}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/listGoals \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.listGoals({});listPlansLists Grammarly's subscription plans with their prices for the visitor's country.Arguments
None.
Returns
GrammarlyPlansinterface GrammarlyPlan { id: number; title: string; description: string; /** The plan's price for the visitor's country, in `currency`. */ price: number; currency: string; periodMonths: number; } interface GrammarlyPlans { /** Whether the anonymous visitor's session already carries a paid plan, always false. */ isPremium: boolean; /** The country the prices are quoted for, e.g. "US". */ countryCode: string; plans: GrammarlyPlan[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/listPlansCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.listPlans();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.listPlans())curl -X POST https://api.bowmark.ai/v1/sites/grammarly/listPlans \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.listPlans();paraphraseRewrites text in different words with the same meaning, logged out (up to 625 characters).Arguments
Name Type Required textstringThe text to reword, up to 625 characters (the logged-out tool's own cap).Yes Returns
Paraphrasedinterface Paraphrased { /** The reworded text: same meaning, same number of sentences. */ text: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/paraphraseCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.paraphrase({ text: "The weather was really bad so we decided to stay at home and watch movies all day." });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.paraphrase({ "text": "The weather was really bad so we decided to stay at home and watch movies all day." }))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/paraphrase \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"text":"The weather was really bad so we decided to stay at home and watch movies all day."}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.paraphrase({ text: "The weather was really bad so we decided to stay at home and watch movies all day." });
addToDictionaryAdds a word or phrase to the signed-in account's personal dictionary so the checker stops flagging it.Arguments
Name Type Required wordstringOne word or short phrase, exactly as it should be accepted.Yes Returns
DictionaryAddedinterface DictionaryAdded { word: string; /** false when Grammarly did not flag the word at all (already accepted or already in the dictionary). */ added: boolean; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/addToDictionaryCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.addToDictionary({ word: "<word>" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.addToDictionary({"word": "<word>"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/addToDictionary \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"word":"<word>"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.addToDictionary({ word: "<word>" });createDocumentSaves a new document (title and text) in the signed-in account, the way app.grammarly.com/ddocs/new does.Arguments
Name Type Required textstringThe document's text.Yes titlestringThe title shown in the documents list.No Returns
GrammarlyDocumentinterface GrammarlyDocument { id: string; title: string; /** The opening of the document's text. */ preview: string; /** Open issues the checker counted. */ errors: number; size: number; /** The uploaded file's name, or null for a document typed in the editor. */ originalFilename: string | null; createdAt: string | null; updatedAt: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/createDocumentCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.createDocument({ text: "<text>" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.createDocument({"text": "<text>"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/createDocument \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"text":"<text>"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.createDocument({ text: "<text>" });deleteDocumentDeletes one saved document from the signed-in account, by the id listDocuments returns.Arguments
Name Type Required idstringThe id listDocuments or createDocument returned.Yes Returns
DeletedDocumentinterface DeletedDocument { id: string; /** Always true: the document went to the trash (listDocuments({ trash: true })). */ trashed: boolean; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/deleteDocumentCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.deleteDocument({ id: "<id>" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.deleteDocument({"id": "<id>"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/deleteDocument \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"id":"<id>"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.deleteDocument({ id: "<id>" });removeFromDictionaryRemoves a word or phrase from the signed-in account's personal dictionary.Arguments
Name Type Required wordstringOne word or short phrase, exactly as it should be accepted.Yes Returns
DictionaryRemovedinterface DictionaryRemoved { word: string; removed: boolean; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/removeFromDictionaryCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.removeFromDictionary({ word: "<word>" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.removeFromDictionary({"word": "<word>"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/removeFromDictionary \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"word":"<word>"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.removeFromDictionary({ word: "<word>" });setGoalsSets the writing goals (goals, domain, audience, formality, dialect) on one saved document in the caller's own account, then reads the document back…Arguments
Name Type Required documentIdstringThe document id listDocuments returned.Yes goalsstring[]Replaces the document's goals.No domainstringAcademic · business · general · mail · casual · creative.No audiencestringGeneral · knowledgeable · expert.No formalitystringInformal · neutral · formal.No dialectstringAmerican · british · canadian · australian · indian.No Returns
GoalsSetinterface GoalsSnapshot { goals: string[]; domain: string | null; audience: string | null; formality: string | null; dialect: string | null; } interface GoalsSet extends GoalsSnapshot { documentId: string; /** What the document carried before this call. Pass it back to restore the goals. */ previous: GoalsSnapshot; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/grammarly/setGoalsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.grammarly.setGoals({ documentId: "<documentId>" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.grammarly.setGoals({"documentId": "<documentId>"}))curl -X POST https://api.bowmark.ai/v1/sites/grammarly/setGoals \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"documentId":"<documentId>"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.grammarly.setGoals({ documentId: "<documentId>" });
Use it from code
- Node
npm i @bowmark/web - Python
pip install bowmark-web - HTTP
POST /v1/sites/grammarly/<function> - OpenAPI
/v1/sites/grammarly/openapi.json