@dataxamas/tracker-js 1.0.1 → 2.1.0

package/README.md

@@ -1,6 +1,6 @@
 # Dataxamas SDK
 
-Official JavaScript/TypeScript SDK for the [Dataxamas](https://dataxamas.izakdvlpr.com) analytics platform.
+Official JavaScript/TypeScript SDK for the [Dataxamas](https://dataxamas.com) analytics platform.
 
 One package, three capabilities:
 
@@ -96,7 +96,7 @@ const client = new Dataxamas(options);
 | Option | Type | Default | Description |
 | --- | --- | --- | --- |
 | `apiKey` | `string` | — | **Required.** Website API key (`dx_live_`) used by logs, errors and tracking. |
-| `baseUrl` | `string` | `https://dataxamas.izakdvlpr.com` | API base URL. |
+| `baseUrl` | `string` | `https://dataxamas.com` | API base URL. |
 | `enableLogs` | `boolean` | `true` | Enable the logs channel. |
 | `enableErrorTracking` | `boolean` | `true` | Enable exception capture. |
 | `captureGlobalErrors` | `boolean` | `true` | Auto-install global error/rejection handlers. |
@@ -188,7 +188,7 @@ tracking.start();
 | Option | Type | Default | Description |
 | --- | --- | --- | --- |
 | `apiKey` | `string` | — | **Required.** Website API key (`dx_live_`). |
-| `baseUrl` | `string` | `https://dataxamas.izakdvlpr.com` | API base URL. |
+| `baseUrl` | `string` | `https://dataxamas.com` | API base URL. |
 | `replayEnabled` | `boolean` | `true` | Allow session replay (gated by server config). |
 | `heatMapEnabled` | `boolean` | `true` | Collect heat-map coordinates on clicks. |
 | `allowLocalhost` | `boolean` | `false` | Allow tracking on `localhost`/`127.0.0.1`. |
@@ -249,7 +249,7 @@ Exported for use in your own typings:
 
 ## Notes
 
-- **Default endpoint:** `https://dataxamas.izakdvlpr.com`. Override with `baseUrl`.
+- **Default endpoint:** `https://dataxamas.com`. Override with `baseUrl`.
 - **Session storage key:** `localStorage["@dataxamas/session"]`.
 - **Event flush spacing:** 4 seconds between queued analytics events.
 - **Replay masking classes:** `dx-mask` (text) and `dx-block` (element); all inputs masked by default.

package/dist/index.cjs

@@ -1,9 +1,116 @@
 "use strict";Object.defineProperty(exports, "__esModule", {value: true}); function _interopRequireWildcard(obj) { if (obj && obj.__esModule) { return obj; } else { var newObj = {}; if (obj != null) { for (var key in obj) { if (Object.prototype.hasOwnProperty.call(obj, key)) { newObj[key] = obj[key]; } } } newObj.default = obj; return newObj; } } function _nullishCoalesce(lhs, rhsFn) { if (lhs != null) { return lhs; } else { return rhsFn(); } } function _optionalChain(ops) { let lastAccessLHS = undefined; let value = ops[0]; let i = 1; while (i < ops.length) { const op = ops[i]; const fn = ops[i + 1]; i += 2; if ((op === 'optionalAccess' || op === 'optionalCall') && value == null) { return undefined; } if (op === 'access' || op === 'optionalAccess') { lastAccessLHS = value; value = fn(value); } else if (op === 'call' || op === 'optionalCall') { value = fn((...args) => value.call(lastAccessLHS, ...args)); lastAccessLHS = undefined; } } return value; }// src/constants.ts
-var BASE_URL = "https://dataxamas.izakdvlpr.com";
+var BASE_URL = "https://dataxamas.com";
 
 // src/utils.ts
 var isBrowser = () => typeof window !== "undefined" && typeof document !== "undefined";
 
+// src/crawlers.ts
+var CRAWLER_UA = /ChatGPT-User|GPTBot|OAI-SearchBot|Claude-User|Claude-SearchBot|ClaudeBot|Claude-Web|anthropic-ai|Perplexity|Gemini-Deep-Research|Googlebot|Google-InspectionTool|Storebot-Google|Google-CloudVertexBot|bingbot|BingPreview|adidxbot|MistralAI-User|DuckAssistBot|DuckDuckBot|meta-external|Applebot|Amazonbot|Bytespider|CCBot|cohere|YandexBot/i;
+var CRAWLER_FILES = /\/(robots\.txt|llms(-full)?\.txt|sitemap[^/]*\.xml)$|\.md$/i;
+var STATIC_ASSET = /\.(js|mjs|css|map|json|png|jpe?g|gif|webp|avif|svg|ico|woff2?|ttf|otf|eot|mp4|webm|mp3|wav|pdf|zip|txt|xml)$/i;
+var IGNORED_PREFIXES = ["/_next/", "/assets/", "/static/", "/api/", "/_build/", "/@vite/", "/node_modules/"];
+function isCrawlerUserAgent(userAgent) {
+  return Boolean(userAgent) && CRAWLER_UA.test(userAgent);
+}
+function shouldTrackPath(pathname) {
+  if (CRAWLER_FILES.test(pathname)) return true;
+  if (IGNORED_PREFIXES.some((prefix) => pathname.startsWith(prefix))) return false;
+  return !STATIC_ASSET.test(pathname);
+}
+function ipFromHeaders(get) {
+  const forwarded = get("x-forwarded-for");
+  if (forwarded) return _optionalChain([forwarded, 'access', _2 => _2.split, 'call', _3 => _3(","), 'access', _4 => _4[0], 'optionalAccess', _5 => _5.trim, 'call', _6 => _6()]) || null;
+  return _nullishCoalesce(_nullishCoalesce(get("cf-connecting-ip"), () => ( get("x-real-ip"))), () => ( null));
+}
+var DataxamasCrawlers = class {
+  constructor(config) {
+    this.config = config;
+  }
+  /**
+   * Reports a fetch-style request if it came from a known crawler. Do not await it: call it and
+   * return your response. The report is fire-and-forget and swallows every network error.
+   */
+  trackRequest(request, options = {}) {
+    const userAgent = request.headers.get("user-agent");
+    if (!userAgent) return;
+    const promise = this.track({
+      url: request.url,
+      userAgent,
+      ip: options.ip !== void 0 ? options.ip : ipFromHeaders((name) => request.headers.get(name)),
+      statusCode: options.statusCode
+    });
+    _optionalChain([options, 'access', _7 => _7.waitUntil, 'optionalCall', _8 => _8(promise)]);
+  }
+  /** Lower-level variant for servers without a `Request` object. Resolves once sent; never rejects. */
+  track(hit) {
+    if (!this.config.enabled || isBrowser() || !isCrawlerUserAgent(hit.userAgent)) {
+      return Promise.resolve();
+    }
+    let url;
+    try {
+      url = new URL(hit.url);
+    } catch (e2) {
+      return Promise.resolve();
+    }
+    if (!shouldTrackPath(url.pathname)) return Promise.resolve();
+    url.search = "";
+    url.hash = "";
+    return fetch(`${this.config.baseUrl}/api/v1/crawls`, {
+      method: "POST",
+      headers: {
+        Authorization: `Bearer ${this.config.apiKey}`,
+        "Content-Type": "application/json"
+      },
+      body: JSON.stringify({
+        url: url.toString(),
+        userAgent: hit.userAgent,
+        ip: _nullishCoalesce(hit.ip, () => ( void 0)),
+        statusCode: hit.statusCode
+      }),
+      signal: AbortSignal.timeout(_nullishCoalesce(this.config.timeoutMs, () => ( 2e3)))
+    }).then(() => void 0).catch(() => void 0);
+  }
+};
+function withCrawlerTracking(crawlers, handler) {
+  return async (request, ...args) => {
+    const response = await handler(request, ...args);
+    crawlers.trackRequest(request, { statusCode: response.status, waitUntil: findWaitUntil(args) });
+    return response;
+  };
+}
+function findWaitUntil(args) {
+  for (const arg of args) {
+    if (arg && typeof arg === "object" && "waitUntil" in arg && typeof arg.waitUntil === "function") {
+      const waitUntil = arg.waitUntil;
+      return (promise) => waitUntil.call(arg, promise);
+    }
+  }
+  return void 0;
+}
+function headerValue(headers, name) {
+  const value = headers[name];
+  return _nullishCoalesce((Array.isArray(value) ? value[0] : value), () => ( null));
+}
+function createExpressCrawlerMiddleware(crawlers) {
+  return (request, response, next) => {
+    const userAgent = headerValue(request.headers, "user-agent");
+    if (isCrawlerUserAgent(userAgent)) {
+      const host = _nullishCoalesce(headerValue(request.headers, "x-forwarded-host"), () => ( headerValue(request.headers, "host")));
+      const protocol = _nullishCoalesce(request.protocol, () => ( "https"));
+      const path = _nullishCoalesce(_nullishCoalesce(request.originalUrl, () => ( request.url)), () => ( "/"));
+      response.once("finish", () => {
+        void crawlers.track({
+          url: `${protocol}://${host}${path}`,
+          userAgent,
+          ip: _nullishCoalesce(_nullishCoalesce(_nullishCoalesce(request.ip, () => ( ipFromHeaders((name) => headerValue(request.headers, name)))), () => ( _optionalChain([request, 'access', _9 => _9.socket, 'optionalAccess', _10 => _10.remoteAddress]))), () => ( null)),
+          statusCode: response.statusCode
+        });
+      });
+    }
+    next();
+  };
+}
+
 // src/exceptions.ts
 function normalizeError(error) {
   if (error instanceof Error) {
@@ -14,7 +121,7 @@ function normalizeError(error) {
   }
   try {
     return { name: "Error", message: JSON.stringify(error) };
-  } catch (e2) {
+  } catch (e3) {
     return { name: "Error", message: String(error) };
   }
 }
@@ -70,7 +177,7 @@ var DataxamasExceptions = class {
     }).then(async (response) => {
       if (response.ok) return;
       const body = await response.json().catch(() => null);
-      const errorMessage = _nullishCoalesce(_optionalChain([body, 'optionalAccess', _2 => _2.error, 'optionalAccess', _3 => _3.message]), () => ( `Request failed with status ${response.status}`));
+      const errorMessage = _nullishCoalesce(_optionalChain([body, 'optionalAccess', _11 => _11.error, 'optionalAccess', _12 => _12.message]), () => ( `Request failed with status ${response.status}`));
       console.error(`[Dataxamas] ${errorMessage}`);
     }).catch(() => {
     });
@@ -108,7 +215,7 @@ function formatData(data) {
   if (data === void 0) return "";
   try {
     return ` ${JSON.stringify(data)}`;
-  } catch (e3) {
+  } catch (e4) {
     return ` ${String(data)}`;
   }
 }
@@ -120,7 +227,7 @@ var nodeDestination = {
       const color = _nullishCoalesce(ANSI[level], () => ( ""));
       process.stdout.write(`${color}[${level}]${RESET} ${_nullishCoalesce(record.msg, () => ( ""))}${formatData(record.data)}
 `);
-    } catch (e4) {
+    } catch (e5) {
       process.stdout.write(line);
     }
   }
@@ -197,7 +304,7 @@ var DataxamasLogs = class {
     }).then(async (response) => {
       if (response.ok) return;
       const body = await response.json().catch(() => null);
-      const errorMessage = _nullishCoalesce(_optionalChain([body, 'optionalAccess', _4 => _4.error, 'optionalAccess', _5 => _5.message]), () => ( `Request failed with status ${response.status}`));
+      const errorMessage = _nullishCoalesce(_optionalChain([body, 'optionalAccess', _13 => _13.error, 'optionalAccess', _14 => _14.message]), () => ( `Request failed with status ${response.status}`));
       this.logger.error(`[Dataxamas] ${errorMessage}`);
     }).catch((error) => {
       this.logger.error(`[Dataxamas] Failed to send log: ${error}`);
@@ -358,10 +465,10 @@ var DataxamasTracking = class {
     };
   }
   safeText(t) {
-    const tag = _optionalChain([t, 'access', _6 => _6.tagName, 'optionalAccess', _7 => _7.toLowerCase, 'call', _8 => _8()]);
+    const tag = _optionalChain([t, 'access', _15 => _15.tagName, 'optionalAccess', _16 => _16.toLowerCase, 'call', _17 => _17()]);
     if (tag === "input" || tag === "textarea" || tag === "select" || tag === "option") return void 0;
     if (t.isContentEditable) return void 0;
-    return _optionalChain([t, 'access', _9 => _9.textContent, 'optionalAccess', _10 => _10.trim, 'call', _11 => _11(), 'access', _12 => _12.substring, 'call', _13 => _13(0, 50)]);
+    return _optionalChain([t, 'access', _18 => _18.textContent, 'optionalAccess', _19 => _19.trim, 'call', _20 => _20(), 'access', _21 => _21.substring, 'call', _22 => _22(0, 50)]);
   }
   safeString(value, max) {
     if (typeof value !== "string") return void 0;
@@ -382,11 +489,11 @@ var DataxamasTracking = class {
         this.sendEvent({
           name: "CLICK",
           data: {
-            tagName: _optionalChain([t, 'access', _14 => _14.tagName, 'optionalAccess', _15 => _15.toLowerCase, 'call', _16 => _16()]),
+            tagName: _optionalChain([t, 'access', _23 => _23.tagName, 'optionalAccess', _24 => _24.toLowerCase, 'call', _25 => _25()]),
             text: this.safeText(t),
             href: this.safeString(t.href, 2048),
             id: this.safeString(t.id, 128),
-            classes: this.safeString(typeof t.className === "string" ? t.className : _optionalChain([t, 'access', _17 => _17.getAttribute, 'optionalCall', _18 => _18("class")]), 256),
+            classes: this.safeString(typeof t.className === "string" ? t.className : _optionalChain([t, 'access', _26 => _26.getAttribute, 'optionalCall', _27 => _27("class")]), 256),
             heatMap: this.buildHeatMap(e)
           }
         });
@@ -503,7 +610,7 @@ var DataxamasTracking = class {
       });
       if (!res.ok) return;
       const body = await res.json().catch(() => null);
-      if (_optionalChain([body, 'optionalAccess', _19 => _19.stop])) {
+      if (_optionalChain([body, 'optionalAccess', _28 => _28.stop])) {
         this.replayHalted = true;
         this.stopReplayRecorder(false);
       }
@@ -651,6 +758,11 @@ var Dataxamas = class {
       heatMapEnabled: _nullishCoalesce(options.heatMapEnabled, () => ( true)),
       allowLocalhost: _nullishCoalesce(options.allowLocalhost, () => ( false))
     });
+    this.crawlers = new DataxamasCrawlers({
+      apiKey: options.apiKey,
+      baseUrl,
+      enabled: _nullishCoalesce(options.enableCrawlerTracking, () => ( true))
+    });
   }
 };
 
@@ -669,4 +781,8 @@ var DataxamasApiError = class extends Error {
 
 
 
-exports.Dataxamas = Dataxamas; exports.DataxamasApiError = DataxamasApiError; exports.DataxamasExceptions = DataxamasExceptions; exports.DataxamasLogs = DataxamasLogs; exports.DataxamasTracking = DataxamasTracking;
+
+
+
+
+exports.Dataxamas = Dataxamas; exports.DataxamasApiError = DataxamasApiError; exports.DataxamasCrawlers = DataxamasCrawlers; exports.DataxamasExceptions = DataxamasExceptions; exports.DataxamasLogs = DataxamasLogs; exports.DataxamasTracking = DataxamasTracking; exports.createExpressCrawlerMiddleware = createExpressCrawlerMiddleware; exports.isCrawlerUserAgent = isCrawlerUserAgent; exports.withCrawlerTracking = withCrawlerTracking;

package/dist/index.d.cts

@@ -1,3 +1,64 @@
+interface CrawlersConfig {
+    apiKey: string;
+    baseUrl: string;
+    enabled: boolean;
+    /** Abort the report after this long. The crawler's response never waits on it either way. */
+    timeoutMs?: number;
+}
+interface CrawlHit {
+    /** Absolute URL the crawler requested. The query string is dropped before sending. */
+    url: string;
+    userAgent: string;
+    /** Crawler's source IP, used to verify it against the provider's published ranges. */
+    ip?: string | null;
+    statusCode?: number;
+}
+interface TrackCrawlOptions {
+    statusCode?: number;
+    /** Overrides the IP read from `x-forwarded-for` / `x-real-ip` / `cf-connecting-ip`. */
+    ip?: string | null;
+    /** Keeps the runtime alive until the report is sent (Workers, Vercel, Next `event.waitUntil`). */
+    waitUntil?: (promise: Promise<unknown>) => void;
+}
+declare function isCrawlerUserAgent(userAgent: string | null | undefined): boolean;
+declare class DataxamasCrawlers {
+    private readonly config;
+    constructor(config: CrawlersConfig);
+    /**
+     * Reports a fetch-style request if it came from a known crawler. Do not await it: call it and
+     * return your response. The report is fire-and-forget and swallows every network error.
+     */
+    trackRequest(request: Request, options?: TrackCrawlOptions): void;
+    /** Lower-level variant for servers without a `Request` object. Resolves once sent; never rejects. */
+    track(hit: CrawlHit): Promise<void>;
+}
+/**
+ * Wraps a fetch-style handler (Cloudflare Workers, Hono, TanStack Start, Bun.serve) so every
+ * crawler request is reported with its final status code. Any argument carrying `waitUntil`
+ * (the Workers `ctx`, Next's `event`) is used to keep the runtime alive for the report.
+ */
+declare function withCrawlerTracking<Args extends unknown[]>(crawlers: DataxamasCrawlers, handler: (request: Request, ...args: Args) => Response | Promise<Response>): (request: Request, ...args: Args) => Promise<Response>;
+interface ExpressLikeRequest {
+    headers: Record<string, string | string[] | undefined>;
+    originalUrl?: string;
+    url?: string;
+    protocol?: string;
+    ip?: string;
+    socket?: {
+        remoteAddress?: string;
+    };
+}
+interface ExpressLikeResponse {
+    statusCode: number;
+    once(event: "finish", listener: () => void): unknown;
+}
+/**
+ * Express / Connect middleware. It reports after the response finishes, so the status code is
+ * the real one, and never delays `next()`. With `app.set("trust proxy", ...)` configured,
+ * `req.ip` already holds the client IP and is preferred over the raw headers.
+ */
+declare function createExpressCrawlerMiddleware(crawlers: DataxamasCrawlers): (request: ExpressLikeRequest, response: ExpressLikeResponse, next: () => void) => void;
+
 type ErrorSource = "CLIENT" | "SERVER";
 type ErrorLevel = "ERROR" | "WARNING" | "FATAL";
 interface ExceptionsConfig {
@@ -122,6 +183,7 @@ interface DataxamasOptions {
     release?: string;
     enableLogs?: boolean;
     enableErrorTracking?: boolean;
+    enableCrawlerTracking?: boolean;
     captureGlobalErrors?: boolean;
     replayEnabled?: boolean;
     heatMapEnabled?: boolean;
@@ -130,6 +192,7 @@ declare class Dataxamas {
     readonly logs: DataxamasLogs;
     readonly exceptions: DataxamasExceptions;
     readonly tracking: DataxamasTracking;
+    readonly crawlers: DataxamasCrawlers;
     constructor(options: DataxamasOptions);
 }
 
@@ -139,4 +202,4 @@ declare class DataxamasApiError extends Error {
     constructor(code: string, message: string, status: number);
 }
 
-export { Dataxamas, DataxamasApiError, DataxamasExceptions, DataxamasLogs, type DataxamasOptions, DataxamasTracking, type DataxamasTrackingOptions, type ErrorLevel, type ErrorSource, type ExceptionsConfig, type LogLevel, type LogsConfig };
+export { type CrawlHit, type CrawlersConfig, Dataxamas, DataxamasApiError, DataxamasCrawlers, DataxamasExceptions, DataxamasLogs, type DataxamasOptions, DataxamasTracking, type DataxamasTrackingOptions, type ErrorLevel, type ErrorSource, type ExceptionsConfig, type LogLevel, type LogsConfig, type TrackCrawlOptions, createExpressCrawlerMiddleware, isCrawlerUserAgent, withCrawlerTracking };

package/dist/index.d.ts

@@ -1,3 +1,64 @@
+interface CrawlersConfig {
+    apiKey: string;
+    baseUrl: string;
+    enabled: boolean;
+    /** Abort the report after this long. The crawler's response never waits on it either way. */
+    timeoutMs?: number;
+}
+interface CrawlHit {
+    /** Absolute URL the crawler requested. The query string is dropped before sending. */
+    url: string;
+    userAgent: string;
+    /** Crawler's source IP, used to verify it against the provider's published ranges. */
+    ip?: string | null;
+    statusCode?: number;
+}
+interface TrackCrawlOptions {
+    statusCode?: number;
+    /** Overrides the IP read from `x-forwarded-for` / `x-real-ip` / `cf-connecting-ip`. */
+    ip?: string | null;
+    /** Keeps the runtime alive until the report is sent (Workers, Vercel, Next `event.waitUntil`). */
+    waitUntil?: (promise: Promise<unknown>) => void;
+}
+declare function isCrawlerUserAgent(userAgent: string | null | undefined): boolean;
+declare class DataxamasCrawlers {
+    private readonly config;
+    constructor(config: CrawlersConfig);
+    /**
+     * Reports a fetch-style request if it came from a known crawler. Do not await it: call it and
+     * return your response. The report is fire-and-forget and swallows every network error.
+     */
+    trackRequest(request: Request, options?: TrackCrawlOptions): void;
+    /** Lower-level variant for servers without a `Request` object. Resolves once sent; never rejects. */
+    track(hit: CrawlHit): Promise<void>;
+}
+/**
+ * Wraps a fetch-style handler (Cloudflare Workers, Hono, TanStack Start, Bun.serve) so every
+ * crawler request is reported with its final status code. Any argument carrying `waitUntil`
+ * (the Workers `ctx`, Next's `event`) is used to keep the runtime alive for the report.
+ */
+declare function withCrawlerTracking<Args extends unknown[]>(crawlers: DataxamasCrawlers, handler: (request: Request, ...args: Args) => Response | Promise<Response>): (request: Request, ...args: Args) => Promise<Response>;
+interface ExpressLikeRequest {
+    headers: Record<string, string | string[] | undefined>;
+    originalUrl?: string;
+    url?: string;
+    protocol?: string;
+    ip?: string;
+    socket?: {
+        remoteAddress?: string;
+    };
+}
+interface ExpressLikeResponse {
+    statusCode: number;
+    once(event: "finish", listener: () => void): unknown;
+}
+/**
+ * Express / Connect middleware. It reports after the response finishes, so the status code is
+ * the real one, and never delays `next()`. With `app.set("trust proxy", ...)` configured,
+ * `req.ip` already holds the client IP and is preferred over the raw headers.
+ */
+declare function createExpressCrawlerMiddleware(crawlers: DataxamasCrawlers): (request: ExpressLikeRequest, response: ExpressLikeResponse, next: () => void) => void;
+
 type ErrorSource = "CLIENT" | "SERVER";
 type ErrorLevel = "ERROR" | "WARNING" | "FATAL";
 interface ExceptionsConfig {
@@ -122,6 +183,7 @@ interface DataxamasOptions {
     release?: string;
     enableLogs?: boolean;
     enableErrorTracking?: boolean;
+    enableCrawlerTracking?: boolean;
     captureGlobalErrors?: boolean;
     replayEnabled?: boolean;
     heatMapEnabled?: boolean;
@@ -130,6 +192,7 @@ declare class Dataxamas {
     readonly logs: DataxamasLogs;
     readonly exceptions: DataxamasExceptions;
     readonly tracking: DataxamasTracking;
+    readonly crawlers: DataxamasCrawlers;
     constructor(options: DataxamasOptions);
 }
 
@@ -139,4 +202,4 @@ declare class DataxamasApiError extends Error {
     constructor(code: string, message: string, status: number);
 }
 
-export { Dataxamas, DataxamasApiError, DataxamasExceptions, DataxamasLogs, type DataxamasOptions, DataxamasTracking, type DataxamasTrackingOptions, type ErrorLevel, type ErrorSource, type ExceptionsConfig, type LogLevel, type LogsConfig };
+export { type CrawlHit, type CrawlersConfig, Dataxamas, DataxamasApiError, DataxamasCrawlers, DataxamasExceptions, DataxamasLogs, type DataxamasOptions, DataxamasTracking, type DataxamasTrackingOptions, type ErrorLevel, type ErrorSource, type ExceptionsConfig, type LogLevel, type LogsConfig, type TrackCrawlOptions, createExpressCrawlerMiddleware, isCrawlerUserAgent, withCrawlerTracking };

package/dist/index.js

@@ -1,9 +1,116 @@
 // src/constants.ts
-var BASE_URL = "https://dataxamas.izakdvlpr.com";
+var BASE_URL = "https://dataxamas.com";
 
 // src/utils.ts
 var isBrowser = () => typeof window !== "undefined" && typeof document !== "undefined";
 
+// src/crawlers.ts
+var CRAWLER_UA = /ChatGPT-User|GPTBot|OAI-SearchBot|Claude-User|Claude-SearchBot|ClaudeBot|Claude-Web|anthropic-ai|Perplexity|Gemini-Deep-Research|Googlebot|Google-InspectionTool|Storebot-Google|Google-CloudVertexBot|bingbot|BingPreview|adidxbot|MistralAI-User|DuckAssistBot|DuckDuckBot|meta-external|Applebot|Amazonbot|Bytespider|CCBot|cohere|YandexBot/i;
+var CRAWLER_FILES = /\/(robots\.txt|llms(-full)?\.txt|sitemap[^/]*\.xml)$|\.md$/i;
+var STATIC_ASSET = /\.(js|mjs|css|map|json|png|jpe?g|gif|webp|avif|svg|ico|woff2?|ttf|otf|eot|mp4|webm|mp3|wav|pdf|zip|txt|xml)$/i;
+var IGNORED_PREFIXES = ["/_next/", "/assets/", "/static/", "/api/", "/_build/", "/@vite/", "/node_modules/"];
+function isCrawlerUserAgent(userAgent) {
+  return Boolean(userAgent) && CRAWLER_UA.test(userAgent);
+}
+function shouldTrackPath(pathname) {
+  if (CRAWLER_FILES.test(pathname)) return true;
+  if (IGNORED_PREFIXES.some((prefix) => pathname.startsWith(prefix))) return false;
+  return !STATIC_ASSET.test(pathname);
+}
+function ipFromHeaders(get) {
+  const forwarded = get("x-forwarded-for");
+  if (forwarded) return forwarded.split(",")[0]?.trim() || null;
+  return get("cf-connecting-ip") ?? get("x-real-ip") ?? null;
+}
+var DataxamasCrawlers = class {
+  constructor(config) {
+    this.config = config;
+  }
+  /**
+   * Reports a fetch-style request if it came from a known crawler. Do not await it: call it and
+   * return your response. The report is fire-and-forget and swallows every network error.
+   */
+  trackRequest(request, options = {}) {
+    const userAgent = request.headers.get("user-agent");
+    if (!userAgent) return;
+    const promise = this.track({
+      url: request.url,
+      userAgent,
+      ip: options.ip !== void 0 ? options.ip : ipFromHeaders((name) => request.headers.get(name)),
+      statusCode: options.statusCode
+    });
+    options.waitUntil?.(promise);
+  }
+  /** Lower-level variant for servers without a `Request` object. Resolves once sent; never rejects. */
+  track(hit) {
+    if (!this.config.enabled || isBrowser() || !isCrawlerUserAgent(hit.userAgent)) {
+      return Promise.resolve();
+    }
+    let url;
+    try {
+      url = new URL(hit.url);
+    } catch {
+      return Promise.resolve();
+    }
+    if (!shouldTrackPath(url.pathname)) return Promise.resolve();
+    url.search = "";
+    url.hash = "";
+    return fetch(`${this.config.baseUrl}/api/v1/crawls`, {
+      method: "POST",
+      headers: {
+        Authorization: `Bearer ${this.config.apiKey}`,
+        "Content-Type": "application/json"
+      },
+      body: JSON.stringify({
+        url: url.toString(),
+        userAgent: hit.userAgent,
+        ip: hit.ip ?? void 0,
+        statusCode: hit.statusCode
+      }),
+      signal: AbortSignal.timeout(this.config.timeoutMs ?? 2e3)
+    }).then(() => void 0).catch(() => void 0);
+  }
+};
+function withCrawlerTracking(crawlers, handler) {
+  return async (request, ...args) => {
+    const response = await handler(request, ...args);
+    crawlers.trackRequest(request, { statusCode: response.status, waitUntil: findWaitUntil(args) });
+    return response;
+  };
+}
+function findWaitUntil(args) {
+  for (const arg of args) {
+    if (arg && typeof arg === "object" && "waitUntil" in arg && typeof arg.waitUntil === "function") {
+      const waitUntil = arg.waitUntil;
+      return (promise) => waitUntil.call(arg, promise);
+    }
+  }
+  return void 0;
+}
+function headerValue(headers, name) {
+  const value = headers[name];
+  return (Array.isArray(value) ? value[0] : value) ?? null;
+}
+function createExpressCrawlerMiddleware(crawlers) {
+  return (request, response, next) => {
+    const userAgent = headerValue(request.headers, "user-agent");
+    if (isCrawlerUserAgent(userAgent)) {
+      const host = headerValue(request.headers, "x-forwarded-host") ?? headerValue(request.headers, "host");
+      const protocol = request.protocol ?? "https";
+      const path = request.originalUrl ?? request.url ?? "/";
+      response.once("finish", () => {
+        void crawlers.track({
+          url: `${protocol}://${host}${path}`,
+          userAgent,
+          ip: request.ip ?? ipFromHeaders((name) => headerValue(request.headers, name)) ?? request.socket?.remoteAddress ?? null,
+          statusCode: response.statusCode
+        });
+      });
+    }
+    next();
+  };
+}
+
 // src/exceptions.ts
 function normalizeError(error) {
   if (error instanceof Error) {
@@ -651,6 +758,11 @@ var Dataxamas = class {
       heatMapEnabled: options.heatMapEnabled ?? true,
       allowLocalhost: options.allowLocalhost ?? false
     });
+    this.crawlers = new DataxamasCrawlers({
+      apiKey: options.apiKey,
+      baseUrl,
+      enabled: options.enableCrawlerTracking ?? true
+    });
   }
 };
 
@@ -666,7 +778,11 @@ var DataxamasApiError = class extends Error {
 export {
   Dataxamas,
   DataxamasApiError,
+  DataxamasCrawlers,
   DataxamasExceptions,
   DataxamasLogs,
-  DataxamasTracking
+  DataxamasTracking,
+  createExpressCrawlerMiddleware,
+  isCrawlerUserAgent,
+  withCrawlerTracking
 };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@dataxamas/tracker-js",
-  "version": "1.0.1",
+  "version": "2.1.0",
   "type": "module",
   "description": "Node.js SDK for Dataxamas Analytics API",
   "license": "MIT",