@apify/actors-mcp-server 0.9.11-beta.2 → 0.9.12-beta.0

package/manifest.json

@@ -2,7 +2,7 @@
   "manifest_version": "0.2",
   "name": "apify-mcp-server",
   "display_name": "Apify",
-  "version": "0.9.11",
+  "version": "0.9.12",
   "description": "Extract data from any website with thousands of scrapers, crawlers, and automations on Apify Store.",
   "long_description": "Apify is the world's largest marketplace of tools for web scraping, data extraction, and web automation. You can extract structured data from social media, e-commerce, search engines, maps, travel sites, or any other website.",
   "keywords": [
@@ -88,7 +88,7 @@
       "linux"
     ],
     "runtimes": {
-      "node": ">=22.0.0"
+      "node": ">=18.0.0"
     }
   }
 }

package/package.json

@@ -1,11 +1,11 @@
 {
   "name": "@apify/actors-mcp-server",
-  "version": "0.9.11-beta.2",
+  "version": "0.9.12-beta.0",
   "type": "module",
   "description": "Apify MCP Server",
   "mcpName": "com.apify/apify-mcp-server",
   "engines": {
-    "node": ">=22.0.0"
+    "node": ">=18.0.0"
   },
   "main": "dist/index.js",
   "exports": {
@@ -47,13 +47,11 @@
     "@modelcontextprotocol/sdk": "^1.25.2",
     "@segment/analytics-node": "^2.3.0",
     "@sentry/node": "^10.38.0",
-    "@types/cheerio": "^0.22.35",
     "@types/turndown": "^5.0.5",
     "ajv": "^8.17.1",
     "algoliasearch": "^5.31.0",
     "apify": "^3.4.2",
     "apify-client": "^2.22.1",
-    "cheerio": "^1.1.2",
     "dotenv": "^16.4.7",
     "express": "^4.21.2",
     "mcp-client-capabilities": "^0.0.5",

package/server.json

@@ -6,7 +6,7 @@
     "url": "https://github.com/apify/apify-mcp-server",
     "source": "github"
   },
-  "version": "0.9.11",
+  "version": "0.9.12",
   "remotes": [
     {
       "type": "streamable-http",

package/dist/tools/common/get_html_skeleton.d.ts

@@ -1,3 +0,0 @@
-import type { ToolEntry } from '../../types.js';
-export declare const getHtmlSkeleton: ToolEntry;
-//# sourceMappingURL=get_html_skeleton.d.ts.map

package/dist/tools/common/get_html_skeleton.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"get_html_skeleton.d.ts","sourceRoot":"","sources":["../../../src/tools/common/get_html_skeleton.ts"],"names":[],"mappings":"AAKA,OAAO,KAAK,EAAoB,SAAS,EAAmB,MAAM,gBAAgB,CAAC;AAgCnF,eAAO,MAAM,eAAe,EAAE,SAyFnB,CAAC"}

package/dist/tools/common/get_html_skeleton.js

@@ -1,102 +0,0 @@
-import { z } from 'zod';
-import { ApifyClient } from '../../apify_client.js';
-import { HelperTools, RAG_WEB_BROWSER, TOOL_MAX_OUTPUT_CHARS, TOOL_STATUS } from '../../const.js';
-import { getHtmlSkeletonCache } from '../../state.js';
-import { compileSchema } from '../../utils/ajv.js';
-import { isValidHttpUrl } from '../../utils/generic.js';
-import { stripHtml } from '../../utils/html.js';
-import { buildMCPResponse } from '../../utils/mcp.js';
-const getHtmlSkeletonArgs = z.object({
-    url: z.string()
-        .min(1)
-        .describe('URL of the webpage to retrieve HTML skeleton from.'),
-    enableJavascript: z.boolean()
-        .optional()
-        .default(false)
-        .describe('Whether to enable JavaScript rendering. Enabling this may increase the time taken to retrieve the HTML skeleton.'),
-    chunk: z.number()
-        .optional()
-        .default(1)
-        .describe('Chunk number to retrieve when getting the content. The content is split into chunks to prevent exceeding the maximum tool output length.'),
-});
-export const getHtmlSkeleton = Object.freeze({
-    type: 'internal',
-    name: HelperTools.GET_HTML_SKELETON,
-    description: `Retrieve the HTML skeleton (clean structure) of a webpage by stripping scripts, styles, and non-essential attributes.
-This keeps the core HTML structure, links, images, and data attributes for analysis. Supports optional JavaScript rendering for dynamic pages.
-
-The results will include a chunked HTML skeleton if the content is large. Use the chunk parameter to paginate through the output.
-
-USAGE:
-- Use when you need a clean HTML structure to design selectors or parsers for scraping.
-
-USAGE EXAMPLES:
-- user_input: Get HTML skeleton for https://example.com
-- user_input: Get next chunk of HTML skeleton for https://example.com (chunk=2)`,
-    inputSchema: z.toJSONSchema(getHtmlSkeletonArgs),
-    ajvValidate: compileSchema(z.toJSONSchema(getHtmlSkeletonArgs)),
-    annotations: {
-        title: 'Get HTML skeleton',
-        readOnlyHint: true,
-        destructiveHint: false,
-        idempotentHint: true,
-        openWorldHint: true,
-    },
-    call: async (toolArgs) => {
-        const { args, apifyToken } = toolArgs;
-        const parsed = getHtmlSkeletonArgs.parse(args);
-        if (!isValidHttpUrl(parsed.url)) {
-            return buildMCPResponse({
-                texts: [`The provided URL is not a valid HTTP or HTTPS URL: ${parsed.url}`],
-                isError: true,
-                toolStatus: TOOL_STATUS.SOFT_FAIL,
-            });
-        }
-        // Try to get from cache first
-        let strippedHtml = getHtmlSkeletonCache.get(parsed.url);
-        if (!strippedHtml) {
-            // Not in cache, call the Actor for scraping
-            const client = new ApifyClient({ token: apifyToken });
-            const run = await client.actor(RAG_WEB_BROWSER).call({
-                query: parsed.url,
-                outputFormats: [
-                    'html',
-                ],
-                scrapingTool: parsed.enableJavascript ? 'browser-playwright' : 'raw-http',
-            });
-            const datasetItems = await client.dataset(run.defaultDatasetId).listItems();
-            if (datasetItems.items.length === 0) {
-                return buildMCPResponse({
-                    texts: [`The scraping Actor (${RAG_WEB_BROWSER}) did not return any output for the URL: ${parsed.url}. Please check the Actor run for more details: ${run.id}`],
-                    isError: true,
-                });
-            }
-            const firstItem = datasetItems.items[0];
-            if (firstItem.crawl.httpStatusMessage.toLocaleLowerCase() !== 'ok') {
-                return buildMCPResponse({
-                    texts: [`The scraping Actor (${RAG_WEB_BROWSER}) returned an HTTP status ${firstItem.crawl.httpStatusCode} (${firstItem.crawl.httpStatusMessage}) for the URL: ${parsed.url}. Please check the Actor run for more details: ${run.id}`],
-                    isError: true,
-                });
-            }
-            if (!firstItem.html) {
-                return buildMCPResponse({
-                    texts: [`The scraping Actor (${RAG_WEB_BROWSER}) did not return any HTML content for the URL: ${parsed.url}. Please check the Actor run for more details: ${run.id}`],
-                    isError: true,
-                });
-            }
-            strippedHtml = stripHtml(firstItem.html);
-            getHtmlSkeletonCache.set(parsed.url, strippedHtml);
-        }
-        // Pagination logic
-        const totalLength = strippedHtml.length;
-        const chunkSize = TOOL_MAX_OUTPUT_CHARS;
-        const totalChunks = Math.ceil(totalLength / chunkSize);
-        const startIndex = (parsed.chunk - 1) * chunkSize;
-        const endIndex = Math.min(startIndex + chunkSize, totalLength);
-        const chunkContent = strippedHtml.slice(startIndex, endIndex);
-        const hasNextChunk = parsed.chunk < totalChunks;
-        const chunkInfo = `\n\n--- Chunk ${parsed.chunk} of ${totalChunks} ---\n${hasNextChunk ? `Next chunk: ${parsed.chunk + 1}` : 'End of content'}`;
-        return buildMCPResponse({ texts: [chunkContent + chunkInfo] });
-    },
-});
-//# sourceMappingURL=get_html_skeleton.js.map

package/dist/tools/common/get_html_skeleton.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"get_html_skeleton.js","sourceRoot":"","sources":["../../../src/tools/common/get_html_skeleton.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,OAAO,EAAE,WAAW,EAAE,MAAM,uBAAuB,CAAC;AACpD,OAAO,EAAE,WAAW,EAAE,eAAe,EAAE,qBAAqB,EAAE,WAAW,EAAE,MAAM,gBAAgB,CAAC;AAClG,OAAO,EAAE,oBAAoB,EAAE,MAAM,gBAAgB,CAAC;AAEtD,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AACnD,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,SAAS,EAAE,MAAM,qBAAqB,CAAC;AAChD,OAAO,EAAE,gBAAgB,EAAE,MAAM,oBAAoB,CAAC;AActD,MAAM,mBAAmB,GAAG,CAAC,CAAC,MAAM,CAAC;IACjC,GAAG,EAAE,CAAC,CAAC,MAAM,EAAE;SACV,GAAG,CAAC,CAAC,CAAC;SACN,QAAQ,CAAC,oDAAoD,CAAC;IACnE,gBAAgB,EAAE,CAAC,CAAC,OAAO,EAAE;SACxB,QAAQ,EAAE;SACV,OAAO,CAAC,KAAK,CAAC;SACd,QAAQ,CAAC,kHAAkH,CAAC;IACjI,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE;SACZ,QAAQ,EAAE;SACV,OAAO,CAAC,CAAC,CAAC;SACV,QAAQ,CAAC,0IAA0I,CAAC;CAC5J,CAAC,CAAC;AAEH,MAAM,CAAC,MAAM,eAAe,GAAc,MAAM,CAAC,MAAM,CAAC;IACpD,IAAI,EAAE,UAAU;IAChB,IAAI,EAAE,WAAW,CAAC,iBAAiB;IACnC,WAAW,EAAE;;;;;;;;;;gFAU+D;IAC5E,WAAW,EAAE,CAAC,CAAC,YAAY,CAAC,mBAAmB,CAAoB;IACnE,WAAW,EAAE,aAAa,CAAC,CAAC,CAAC,YAAY,CAAC,mBAAmB,CAAC,CAAC;IAC/D,WAAW,EAAE;QACT,KAAK,EAAE,mBAAmB;QAC1B,YAAY,EAAE,IAAI;QAClB,eAAe,EAAE,KAAK;QACtB,cAAc,EAAE,IAAI;QACpB,aAAa,EAAE,IAAI;KACtB;IACD,IAAI,EAAE,KAAK,EAAE,QAA0B,EAAE,EAAE;QACvC,MAAM,EAAE,IAAI,EAAE,UAAU,EAAE,GAAG,QAAQ,CAAC;QACtC,MAAM,MAAM,GAAG,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAE/C,IAAI,CAAC,cAAc,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;YAC9B,OAAO,gBAAgB,CAAC;gBACpB,KAAK,EAAE,CAAC,sDAAsD,MAAM,CAAC,GAAG,EAAE,CAAC;gBAC3E,OAAO,EAAE,IAAI;gBACb,UAAU,EAAE,WAAW,CAAC,SAAS;aACpC,CAAC,CAAC;QACP,CAAC;QAED,8BAA8B;QAC9B,IAAI,YAAY,GAAG,oBAAoB,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QACxD,IAAI,CAAC,YAAY,EAAE,CAAC;YAChB,4CAA4C;YAC5C,MAAM,MAAM,GAAG,IAAI,WAAW,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,CAAC,CAAC;YAEtD,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC,IAAI,CAAC;gBACjD,KAAK,EAAE,MAAM,CAAC,GAAG;gBACjB,aAAa,EAAE;oBACX,MAAM;iBACT;gBACD,YAAY,EAAE,MAAM,CAAC,gBAAgB,CAAC,CAAC,CAAC,oBAAoB,CAAC,CAAC,CAAC,UAAU;aAC5E,CAAC,CAAC;YAEH,MAAM,YAAY,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,gBAAgB,CAAC,CAAC,SAAS,EAAE,CAAC;YAC5E,IAAI,YAAY,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBAClC,OAAO,gBAAgB,CAAC;oBACpB,KAAK,EAAE,CAAC,uBAAuB,eAAe,4CAA4C,MAAM,CAAC,GAAG,kDAAkD,GAAG,CAAC,EAAE,EAAE,CAAC;oBAC/J,OAAO,EAAE,IAAI;iBAChB,CAAC,CAAC;YACP,CAAC;YAED,MAAM,SAAS,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC,CAA+B,CAAC;YACtE,IAAI,SAAS,CAAC,KAAK,CAAC,iBAAiB,CAAC,iBAAiB,EAAE,KAAK,IAAI,EAAE,CAAC;gBACjE,OAAO,gBAAgB,CAAC;oBACpB,KAAK,EAAE,CAAC,uBAAuB,eAAe,6BAA6B,SAAS,CAAC,KAAK,CAAC,cAAc,KAAK,SAAS,CAAC,KAAK,CAAC,iBAAiB,kBAAkB,MAAM,CAAC,GAAG,kDAAkD,GAAG,CAAC,EAAE,EAAE,CAAC;oBACtO,OAAO,EAAE,IAAI;iBAChB,CAAC,CAAC;YACP,CAAC;YAED,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;gBAClB,OAAO,gBAAgB,CAAC;oBACpB,KAAK,EAAE,CAAC,uBAAuB,eAAe,kDAAkD,MAAM,CAAC,GAAG,kDAAkD,GAAG,CAAC,EAAE,EAAE,CAAC;oBACrK,OAAO,EAAE,IAAI;iBAChB,CAAC,CAAC;YACP,CAAC;YAED,YAAY,GAAG,SAAS,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;YACzC,oBAAoB,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,EAAE,YAAY,CAAC,CAAC;QACvD,CAAC;QAED,mBAAmB;QACnB,MAAM,WAAW,GAAG,YAAY,CAAC,MAAM,CAAC;QACxC,MAAM,SAAS,GAAG,qBAAqB,CAAC;QACxC,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,WAAW,GAAG,SAAS,CAAC,CAAC;QACvD,MAAM,UAAU,GAAG,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,SAAS,CAAC;QAClD,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,UAAU,GAAG,SAAS,EAAE,WAAW,CAAC,CAAC;QAC/D,MAAM,YAAY,GAAG,YAAY,CAAC,KAAK,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;QAC9D,MAAM,YAAY,GAAG,MAAM,CAAC,KAAK,GAAG,WAAW,CAAC;QAEhD,MAAM,SAAS,GAAG,iBAAiB,MAAM,CAAC,KAAK,OAAO,WAAW,SAAS,YAAY,CAAC,CAAC,CAAC,eAAe,MAAM,CAAC,KAAK,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,gBAAgB,EAAE,CAAC;QAEhJ,OAAO,gBAAgB,CAAC,EAAE,KAAK,EAAE,CAAC,YAAY,GAAG,SAAS,CAAC,EAAE,CAAC,CAAC;IACnE,CAAC;CACK,CAAC,CAAC"}

package/dist/utils/html.d.ts

@@ -1,11 +0,0 @@
-/**
- * Strips HTML and keeps only the structure.
- *
- * Removes styles, scripts, and other non-content elements.
- * Collapses whitespace and trims the result.
- * Keeps only href, src, alt, id, class, title, name, data-* attributes.
- * Removes HTML comments and spaces between tags.
- * Removes base64 encoded images.
- */
-export declare function stripHtml(html: string): string;
-//# sourceMappingURL=html.d.ts.map

package/dist/utils/html.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"html.d.ts","sourceRoot":"","sources":["../../src/utils/html.ts"],"names":[],"mappings":"AAWA;;;;;;;;GAQG;AACH,wBAAgB,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAuC9C"}

package/dist/utils/html.js

@@ -1,48 +0,0 @@
-import * as cheerio from 'cheerio';
-/**
- * Strips HTML and keeps only the structure.
- *
- * Removes styles, scripts, and other non-content elements.
- * Collapses whitespace and trims the result.
- * Keeps only href, src, alt, id, class, title, name, data-* attributes.
- * Removes HTML comments and spaces between tags.
- * Removes base64 encoded images.
- */
-export function stripHtml(html) {
-    const $ = cheerio.load(html);
-    // Remove all attributes except href (only on a), src, alt, id, class, title, name, data-*
-    const allowedAttrs = ['href', 'src', 'alt', 'id', 'class', 'title', 'name'];
-    $('*').each((_, element) => {
-        const { attribs } = element;
-        if (attribs) {
-            Object.keys(attribs).forEach((attr) => {
-                if (attr === 'href' && element.tagName !== 'a') {
-                    $(element).removeAttr(attr);
-                }
-                else if (!allowedAttrs.includes(attr) && !attr.startsWith('data-')) {
-                    $(element).removeAttr(attr);
-                }
-            });
-        }
-    });
-    // Remove <style>, <script>, <noscript>, <iframe>, <svg>, <canvas>, <math> tags and their content
-    $('style, script, noscript, iframe, svg, canvas, math').remove();
-    // Remove HTML comments
-    $('*').contents().filter((_, element) => element.type === 'comment').remove();
-    // Remove base64 encoded images
-    $('img[src^="data:image/"]').remove();
-    let result;
-    if (html.trim() === '') {
-        result = '';
-    }
-    else if (html.includes('<html')) {
-        result = $.html();
-    }
-    else {
-        result = $('body').html() || '';
-    }
-    // Collapse multiple spaces into one, remove spaces between tags, and trim
-    result = result.replace(/\s+/g, ' ').replace(/>\s+</g, '><').trim();
-    return result;
-}
-//# sourceMappingURL=html.js.map

package/dist/utils/html.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"html.js","sourceRoot":"","sources":["../../src/utils/html.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,OAAO,MAAM,SAAS,CAAC;AAWnC;;;;;;;;GAQG;AACH,MAAM,UAAU,SAAS,CAAC,IAAY;IAClC,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAE7B,0FAA0F;IAC1F,MAAM,YAAY,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;IAC5E,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,EAAE;QACvB,MAAM,EAAE,OAAO,EAAE,GAAI,OAA8B,CAAC;QACpD,IAAI,OAAO,EAAE,CAAC;YACV,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;gBAClC,IAAI,IAAI,KAAK,MAAM,IAAK,OAA8B,CAAC,OAAO,KAAK,GAAG,EAAE,CAAC;oBACrE,CAAC,CAAC,OAAO,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;gBAChC,CAAC;qBAAM,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;oBACnE,CAAC,CAAC,OAAO,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;gBAChC,CAAC;YACL,CAAC,CAAC,CAAC;QACP,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,iGAAiG;IACjG,CAAC,CAAC,oDAAoD,CAAC,CAAC,MAAM,EAAE,CAAC;IAEjE,uBAAuB;IACvB,CAAC,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,EAAE,CAAE,OAAoB,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,MAAM,EAAE,CAAC;IAE5F,+BAA+B;IAC/B,CAAC,CAAC,yBAAyB,CAAC,CAAC,MAAM,EAAE,CAAC;IAEtC,IAAI,MAAM,CAAC;IACX,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;QACrB,MAAM,GAAG,EAAE,CAAC;IAChB,CAAC;SAAM,IAAI,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,EAAE,CAAC;QAChC,MAAM,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IACtB,CAAC;SAAM,CAAC;QACJ,MAAM,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC;IACpC,CAAC;IAED,0EAA0E;IAC1E,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;IACpE,OAAO,MAAM,CAAC;AAClB,CAAC"}