@webcrawl/mcp 1.3.7 → 1.3.8

package/dist/format.d.ts

@@ -10,4 +10,5 @@ export declare function formatScrape(res: Rec): string;
 export declare function formatParse(res: Rec): string;
 export declare function formatMap(res: Rec): string;
 export declare function formatCrawl(res: Rec): string;
+export declare function formatMonitor(res: Rec): string;
 export {};

package/dist/format.js

@@ -124,3 +124,66 @@ export function formatCrawl(res) {
     })
         .join("\n\n---\n\n");
 }
+function formatMonitorSummaryLine(m, i) {
+    const prefix = i !== undefined ? `${i + 1}. ` : "";
+    const id = str(m.id);
+    const name = str(m.name) || str(m.url) || "(unnamed watch)";
+    const status = str(m.status) || "active";
+    const mode = str(m.mode) || "page";
+    const urls = Array.isArray(m.urls) ? m.urls.map(str).filter(Boolean) : [];
+    const scopeLabel = mode === "site"
+        ? `site (up to ${str(m.maxPages) || "5"} pages)`
+        : mode === "urls"
+            ? `${urls.length || 1} URLs`
+            : "single page";
+    const interval = typeof m.intervalMinutes === "number" ? `every ${m.intervalMinutes}m` : "";
+    const condExpr = str(m.conditionJev) || (typeof m.condition === "string" ? m.condition : "");
+    const goal = condExpr ? `rule: ${condExpr}` : str(m.watchFor) ? `goal: ${str(m.watchFor)}` : "any change";
+    const condMet = typeof m.lastConditionMet === "boolean"
+        ? m.lastConditionMet
+            ? "triggered"
+            : "waiting"
+        : "";
+    const meta = [status, scopeLabel, interval, condMet].filter(Boolean).join(" · ");
+    const diff = str(m.lastDiffSummary);
+    return [
+        `${prefix}${name} [${id}] (${meta})`,
+        `   URL: ${str(m.url)}${urls.length > 1 ? ` (+${urls.length - 1} more)` : ""}`,
+        `   Watch: ${goal}`,
+        diff ? `   Latest: ${diff}` : "",
+    ]
+        .filter(Boolean)
+        .join("\n");
+}
+export function formatMonitor(res) {
+    if (res.deleted === true) {
+        return "Monitor deleted.";
+    }
+    if (Array.isArray(res.monitors)) {
+        const list = res.monitors.filter(isRec);
+        if (!list.length) {
+            return "No active watches.";
+        }
+        return `monitors (${list.length}):\n${list.map((m, i) => formatMonitorSummaryLine(m, i)).join("\n\n")}`;
+    }
+    const m = isRec(res.monitor) ? res.monitor : isRec(res.data) ? res.data : null;
+    if (!m) {
+        return JSON.stringify(res, null, 2);
+    }
+    const check = isRec(res.check) ? res.check : isRec(res.initialCheck) ? res.initialCheck : null;
+    const parts = [formatMonitorSummaryLine(m)];
+    if (check) {
+        const condRes = isRec(check.conditionResult) ? check.conditionResult : null;
+        if (condRes && str(condRes.summary)) {
+            parts.push(`Condition: ${str(condRes.summary)} (met: ${Boolean(condRes.met)})`);
+        }
+        if (str(check.diffSummary)) {
+            parts.push(`Summary: ${str(check.diffSummary)}`);
+        }
+    }
+    const snap = (check && str(check.snapshot)) || str(m.lastSnapshot);
+    if (snap) {
+        parts.push(`Snapshot:\n${truncate(snap, 6_000)}`);
+    }
+    return parts.join("\n\n");
+}

package/dist/server.d.ts

@@ -7,6 +7,12 @@ export interface McpServerOptions {
 }
 export type WebcrawlLikeClient = Pick<WebcrawlClient, "baseUrl" | "search" | "scrape" | "parse" | "run" | "map" | "crawl" | "extract" | "rerank" | "categories"> & {
     targets?: () => Promise<Record<string, unknown>>;
+    monitors?: WebcrawlClient["monitors"];
+    getMonitor?: WebcrawlClient["getMonitor"];
+    createMonitor?: WebcrawlClient["createMonitor"];
+    checkMonitor?: WebcrawlClient["checkMonitor"];
+    updateMonitor?: WebcrawlClient["updateMonitor"];
+    deleteMonitor?: WebcrawlClient["deleteMonitor"];
 };
 export declare function resolveSchemaProfile(env?: NodeJS.ProcessEnv): SchemaProfile;
 /**

package/dist/server.js

@@ -1,7 +1,7 @@
 import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
-import { describeError } from "@webcrawl/client";
+import { WebcrawlError, describeError } from "@webcrawl/client";
 import { z } from "zod";
-import { formatCredits, formatCrawl, formatMap, formatParse, formatScrape, formatSearch, truncate } from "./format.js";
+import { formatCredits, formatCrawl, formatMap, formatMonitor, formatParse, formatScrape, formatSearch, truncate, } from "./format.js";
 const INSTRUCTIONS = `Webcrawl Anything gives you live web & target inspection access (zero-key local engine or webcrawl.app cloud).
 - web_search: find pages or structured results. Set category for flights, hotels, real-estate, shopping, news, local business, security (CVEs), academic or tech. Results are localised to the user.
 - web_scrape: read one URL as clean markdown with target evidence (websites, SPAs, GitHub, ArXiv, llms.txt).
@@ -10,6 +10,7 @@ const INSTRUCTIONS = `Webcrawl Anything gives you live web & target inspection a
 - web_crawl: read several pages of one site (max 25).
 - web_extract: pull structured JSON from one URL with a prompt or schema.
 - web_rerank: order your own documents by relevance to a query.
+- web_monitor: watch a page, list of URLs, or whole website on a schedule with trigger conditions and email/webhook alerts (create, list, get, check, pause, resume, delete).
 - search_categories: list valid categories, subcategories, filters, and supported crawl targets.`;
 const ok = (text) => ({ content: [{ type: "text", text: truncate(text) }] });
 export function resolveSchemaProfile(env = process.env) {
@@ -183,6 +184,102 @@ export function createWebcrawlMcpServer(client, versionOrOpts = "0.1.0") {
         });
         return (lines.length ? lines.join("\n") : JSON.stringify(res, null, 2)) + formatCredits(res);
     }));
+    server.registerTool("web_monitor", {
+        title: "Monitor web pages or sites",
+        description: compact
+            ? "Create, list, inspect, check, pause, resume, or delete scheduled web monitors with trigger conditions."
+            : "Watch a single page, a list of URLs, or a whole website section on a schedule. Supports natural-language `watch_for` focus, explicit `condition` rules (e.g. `price < 50`, `content contains \"In Stock\"`, `added_urls_count > 0`), and email or webhook notifications.",
+        inputSchema: {
+            action: desc(z.enum(["create", "list", "get", "check", "pause", "resume", "delete"]).optional(), "Action to perform (defaults to 'create' when url/urls given, else 'list')"),
+            id: desc(z.string().optional(), "Monitor ID (required for get, check, pause, resume, delete)"),
+            url: desc(z.string().url().optional(), "Page or site URL to watch"),
+            urls: desc(z.array(z.string().url()).optional(), "Multiple URLs to watch when mode is 'urls'"),
+            mode: desc(z.enum(["page", "urls", "site"]).optional(), "Watch scope: page (default), urls, or site"),
+            max_pages: desc(z.number().int().min(1).max(25).optional(), "Max pages to sample when mode is 'site' (default 5)"),
+            name: desc(z.string().optional(), "Optional friendly watch name"),
+            interval_minutes: desc(z.number().int().min(5).max(10080).optional(), "Check interval in minutes (default 60)"),
+            watch_for: desc(z.string().optional(), "What to watch for, e.g. 'Pricing tiers and discounts'"),
+            condition: desc(z.string().optional(), "Optional trigger rule expression, e.g. 'price < 50', 'content contains \"In Stock\"', 'added_urls_count > 0'"),
+            notify_email: desc(z.string().optional(), "Email address to notify when triggered, or 'true' to use the account email"),
+            webhook_url: desc(z.string().url().optional(), "HTTPS webhook URL to POST when triggered"),
+        },
+        annotations: { readOnlyHint: false, openWorldHint: true },
+    }, ({ action, id, url, urls, mode, max_pages, name, interval_minutes, watch_for, condition, notify_email, webhook_url, }) => run(async () => {
+        const effectiveAction = action ?? (url || (urls && urls.length) ? "create" : "list");
+        const requireId = (op) => {
+            if (!id) {
+                throw new WebcrawlError("bad_request", `web_monitor action '${op}' requires 'id'`);
+            }
+            return id;
+        };
+        switch (effectiveAction) {
+            case "list": {
+                if (!client.monitors) {
+                    throw new WebcrawlError("bad_request", "Web monitors require a cloud WebcrawlClient (set WEBCRAWL_API_KEY).");
+                }
+                const res = await client.monitors();
+                return formatMonitor(res) + formatCredits(res);
+            }
+            case "get": {
+                if (!client.getMonitor) {
+                    throw new WebcrawlError("bad_request", "Web monitors require a cloud WebcrawlClient (set WEBCRAWL_API_KEY).");
+                }
+                const res = await client.getMonitor(requireId("get"));
+                return formatMonitor(res) + formatCredits(res);
+            }
+            case "check": {
+                if (!client.checkMonitor) {
+                    throw new WebcrawlError("bad_request", "Web monitors require a cloud WebcrawlClient (set WEBCRAWL_API_KEY).");
+                }
+                const res = await client.checkMonitor(requireId("check"));
+                return formatMonitor(res) + formatCredits(res);
+            }
+            case "pause":
+            case "resume": {
+                if (!client.updateMonitor) {
+                    throw new WebcrawlError("bad_request", "Web monitors require a cloud WebcrawlClient (set WEBCRAWL_API_KEY).");
+                }
+                const res = await client.updateMonitor({
+                    id: requireId(effectiveAction),
+                    status: effectiveAction === "pause" ? "paused" : "active",
+                    ...(name !== undefined ? { name } : {}),
+                    ...(interval_minutes !== undefined ? { intervalMinutes: interval_minutes } : {}),
+                    ...(watch_for !== undefined ? { watchFor: watch_for } : {}),
+                    ...(condition !== undefined ? { condition } : {}),
+                    ...(notify_email !== undefined
+                        ? { notifyEmail: notify_email === "true" ? true : notify_email }
+                        : {}),
+                    ...(webhook_url !== undefined ? { webhookUrl: webhook_url } : {}),
+                });
+                return formatMonitor(res) + formatCredits(res);
+            }
+            case "delete": {
+                if (!client.deleteMonitor) {
+                    throw new WebcrawlError("bad_request", "Web monitors require a cloud WebcrawlClient (set WEBCRAWL_API_KEY).");
+                }
+                const res = await client.deleteMonitor(requireId("delete"));
+                return formatMonitor(res) + formatCredits(res);
+            }
+            case "create": {
+                if (!client.createMonitor) {
+                    throw new WebcrawlError("bad_request", "Web monitors require a cloud WebcrawlClient (set WEBCRAWL_API_KEY).");
+                }
+                const res = await client.createMonitor({
+                    url: url ?? urls?.[0],
+                    urls,
+                    mode: mode ?? (urls && urls.length > 1 ? "urls" : "page"),
+                    maxPages: max_pages,
+                    name,
+                    intervalMinutes: interval_minutes,
+                    watchFor: watch_for,
+                    condition,
+                    notifyEmail: notify_email === "true" ? true : notify_email,
+                    webhookUrl: webhook_url,
+                });
+                return formatMonitor(res) + formatCredits(res);
+            }
+        }
+    }));
     server.registerTool("search_categories", {
         title: "Search categories",
         description: "List the search categories, subcategories and filters web_search supports.",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@webcrawl/mcp",
-  "version": "1.3.7",
+  "version": "1.3.8",
   "description": "MCP server for Webcrawl Anything — zero-key local engine or webcrawl.app cloud (search, scrape, parse PDF/Office/YouTube, map, crawl, extract, rerank)",
   "type": "module",
   "license": "MIT",
@@ -52,7 +52,7 @@
   },
   "dependencies": {
     "@modelcontextprotocol/sdk": "^1.32.2",
-    "@webcrawl/client": "^1.3.7",
+    "@webcrawl/client": "^1.3.8",
     "zod": "^4.0.0"
   },
   "engines": {