@apmantza/greedysearch-pi 1.8.9 → 1.8.10

package/CHANGELOG.md

@@ -2,6 +2,22 @@
 
 ## [Unreleased]
 
+## [1.8.10] — 2026-05-11
+
+### Removed
+
+- **Dead `bing-aria` extractor** (`extractors/bing-aria.mjs`) — Removed the unused ARIA-tree-based Bing Copilot extractor and its `bing2` engine alias. Nothing in the system referenced it (`bing2` was not in `ALL_ENGINES`, not documented in the tool schema, and had no callers).
+
+### Fixed
+
+- **Perplexity sign-in mis-click** (`extractors/consent.mjs`) — `handleVerification` matched any button containing "continue", including "Continue with Google" OAuth buttons on Perplexity sign-in modals. This caused the automation to accidentally open Google/Microsoft login flows. Added explicit exclusions for `sign.in`, `log.in`, `google`, `microsoft`, `apple`, `facebook`, `github`, and `auth` text patterns in both `VERIFY_DETECT_JS` and `VERIFY_RETRY_JS`.
+
+- **Gemini synthesis typing failure** (`extractors/gemini.mjs`) — `document.execCommand('insertText')` silently failed for long synthesis prompts (~8-10k chars), causing the extractor to submit an empty input and wait forever (45s stream + 180s timeout). Replaced with CDP `Input.insertText` + explicit focus click + content-length verification. Now fails fast with a clear error if text doesn't land.
+
+- **Gemini answer extraction — query echo** (`extractors/gemini.mjs`) — When the assistant response copy button hadn't hydrated yet, clicking `buttons[buttons.length - 1]` hit the user's message copy button instead of Gemini's response, returning the query text as the "answer". Added wait for the assistant copy button to appear (2+ buttons on page), plus retry logic that detects exact query-text echo and re-clicks after a settle delay.
+
+- **Bing Copilot Cloudflare auto-bypass** (`extractors/consent.mjs`) — Copilot's Turnstile challenge lives inside a **closed shadow DOM**, invisible to `document.querySelector('iframe')`. Added detection for the queryable host container (`#cf-turnstile`) and hidden response input (`[id^="cf-chl-widget-"]`), returning center coordinates for `humanClickXY`. During visible recovery, the challenge now auto-clicks and resolves transparently.
+
 ## [1.8.9] — 2026-05-11
 
 ### Changed

package/extractors/consent.mjs

@@ -53,6 +53,15 @@ const VERIFY_DETECT_JS = `
     return JSON.stringify({t:'xy',x:r.left+30,y:r.top+r.height/2});
   }
 
+  // --- Cloudflare Turnstile widget inside closed shadow DOM (Copilot, etc.) ---
+  // The iframe is not queryable from main document, but the host container
+  // (#cf-turnstile) and the hidden response input are.
+  var cfTurnstileHost = document.querySelector('#cf-turnstile, [id^="cf-chl-widget-"]');
+  if (cfTurnstileHost) {
+    var r2 = cfTurnstileHost.getBoundingClientRect();
+    return JSON.stringify({t:'xy',x:r2.left+r2.width/2,y:r2.top+r2.height/2});
+  }
+
   // --- Cloudflare challenge page ---
   var cfCheckbox = document.querySelector('#cf-stage input[type="checkbox"], .ctp-checkbox-container input');
   if (cfCheckbox) { cfCheckbox.setAttribute('data-gs-verify','1'); return JSON.stringify({t:'sel',s:'[data-gs-verify="1"]',txt:'cloudflare-checkbox'}); }
@@ -66,11 +75,16 @@ const VERIFY_DETECT_JS = `
   }
 
   // --- Generic verify/continue/proceed buttons (catch-all) ---
+  // IMPORTANT: exclude sign-in / OAuth buttons (e.g. "Continue with Google")
   var btns = Array.from(document.querySelectorAll('button, input[type=submit], a[role=button]'));
   var verify = btns.find(b => {
     var t = (b.innerText?.trim() || b.value || '').toLowerCase();
-    return (t.includes('verify') || t.includes('human') || t.includes('robot') || t.includes('continue') || t.includes('proceed')) &&
+    var isVerifyLike = (t.includes('verify') || t.includes('human') || t.includes('robot') || t.includes('continue') || t.includes('proceed')) &&
            !t.includes('verified') && !document.querySelector('iframe[src*="recaptcha"]');
+    if (!isVerifyLike) return false;
+    // Exclude OAuth / sign-in buttons to prevent accidental login flows
+    var isSignIn = /sign.in|log.in|google|microsoft|apple|facebook|github|auth/i.test(t);
+    return !isSignIn;
   });
   if (verify) { verify.setAttribute('data-gs-verify','1'); return JSON.stringify({t:'sel',s:'[data-gs-verify="1"]',txt:verify.innerText?.trim()||verify.value}); }
 
@@ -95,13 +109,20 @@ const VERIFY_RETRY_JS = `
   var btns = Array.from(document.querySelectorAll('button, input[type=submit], a[role=button]'));
   var btn = btns.find(b => {
     var t = (b.innerText?.trim() || b.value || '').toLowerCase();
-    return t.includes('verify') || t.includes('human') || t.includes('robot') || t.includes('continue') || t.includes('next') || t.includes('submit');
+    var isVerifyLike = t.includes('verify') || t.includes('human') || t.includes('robot') || t.includes('continue') || t.includes('next') || t.includes('submit');
+    if (!isVerifyLike) return false;
+    var isSignIn = /sign.in|log.in|google|microsoft|apple|facebook|github|auth/i.test(t);
+    return !isSignIn;
   });
   if (btn) { btn.setAttribute('data-gs-verify','1'); return JSON.stringify({t:'sel',s:'[data-gs-verify="1"]',txt:btn.innerText?.trim()||btn.value}); }
 
   var cf = document.querySelector('#cf-stage input[type="checkbox"], .cf-turnstile input');
   if (cf) { cf.setAttribute('data-gs-verify','1'); return JSON.stringify({t:'sel',s:'[data-gs-verify="1"]',txt:'turnstile'}); }
 
+  // Cloudflare Turnstile widget inside closed shadow DOM (detected via host container)
+  var cfTurnstileHost = document.querySelector('#cf-turnstile, [id^="cf-chl-widget-"]');
+  if (cfTurnstileHost) { return 'still-verifying'; }
+
   var modal = document.querySelector('[role="dialog"], .b_modal, [class*="verify"]');
   if (modal) {
     var modalBtn = modal.querySelector('button, a[role="button"]');

package/extractors/gemini.mjs

@@ -36,19 +36,30 @@ const GLOBAL_VAR = "__geminiClipboard";
 // ============================================================================
 
 async function typeIntoGemini(tab, text) {
-	await cdp([
+	// 1. Focus the input area via click (more reliable than eval focus for shadow-DOM editors)
+	await cdp(["click", tab, S.input]);
+	await new Promise((r) => setTimeout(r, jitter(200)));
+
+	// 2. Type using CDP Input.insertText (more reliable than document.execCommand)
+	await cdp(["type", tab, text]);
+	await new Promise((r) => setTimeout(r, jitter(300)));
+
+	// 3. Verify the text was actually inserted
+	const inserted = await cdp([
 		"eval",
 		tab,
-		`
-    (function(t) {
-      var el = document.querySelector('${S.input}');
-      if (!el) return false;
-      el.focus();
-      document.execCommand('insertText', false, t);
-      return true;
-    })(${JSON.stringify(text)})
-  `,
+		`(function() {
+			var el = document.querySelector('${S.input}');
+			if (!el) return false;
+			var content = el.innerText || el.textContent || '';
+			return content.trim().length >= ${Math.floor(text.length * 0.8)};
+		})()`,
 	]);
+	if (inserted !== "true") {
+		throw new Error(
+			"Gemini input field did not accept text — input verification failed",
+		);
+	}
 }
 
 async function scrollToBottom(tab) {
@@ -62,9 +73,32 @@ async function scrollToBottom(tab) {
 	]);
 }
 
-async function extractAnswer(tab) {
-	// Click the LAST copy button (assistant's response at the bottom),
-	// not the first (which could be the user's echoed query).
+async function extractAnswer(tab, query = "") {
+	const queryNorm = query.toLowerCase().trim();
+
+	// Wait for the assistant response copy button to appear.
+	// A fresh conversation has 1 copy button (user message); after the
+	// assistant responds there are 2+.  This prevents clicking the user's
+	// copy button before React hydrates the assistant's.
+	let copyReady = false;
+	const copyDeadline = Date.now() + 12000;
+	while (Date.now() < copyDeadline) {
+		const count = await cdp([
+			"eval",
+			tab,
+			`document.querySelectorAll('${S.copyButton}').length`,
+		]);
+		if (parseInt(count, 10) >= 2) {
+			copyReady = true;
+			break;
+		}
+		await new Promise((r) => setTimeout(r, 800));
+	}
+	if (!copyReady) {
+		console.error("[gemini] Warning: assistant copy button did not appear");
+	}
+
+	// Click the LAST copy button (assistant's response at the bottom)
 	await cdp([
 		"eval",
 		tab,
@@ -73,9 +107,32 @@ async function extractAnswer(tab) {
 			buttons[buttons.length - 1]?.click();
 		})()`,
 	]);
-	await new Promise((r) => setTimeout(r, 400));
+	await new Promise((r) => setTimeout(r, 600));
+
+	let answer = await cdp(["eval", tab, `window.${GLOBAL_VAR} || ''`]);
+
+	// Retry once if clipboard contains the user's query instead of the response.
+	// This can happen when the assistant response hasn't rendered its copy button yet.
+	if (
+		answer &&
+		queryNorm &&
+		(answer.toLowerCase().trim() === queryNorm ||
+			answer.trim().length < queryNorm.length)
+	) {
+		console.error("[gemini] Clipboard echoed query, retrying in 2s...");
+		await new Promise((r) => setTimeout(r, 2000));
+		await cdp([
+			"eval",
+			tab,
+			`(() => {
+				const buttons = document.querySelectorAll('${S.copyButton}');
+				buttons[buttons.length - 1]?.click();
+			})()`,
+		]);
+		await new Promise((r) => setTimeout(r, 600));
+		answer = await cdp(["eval", tab, `window.${GLOBAL_VAR} || ''`]);
+	}
 
-	const answer = await cdp(["eval", tab, `window.${GLOBAL_VAR} || ''`]);
 	if (!answer) throw new Error("Clipboard interceptor returned empty text");
 
 	const sources = parseSourcesFromMarkdown(answer);
@@ -130,7 +187,7 @@ async function main() {
 			clearInterval(scrollInterval);
 		}
 
-		const { answer, sources } = await extractAnswer(tab);
+		const { answer, sources } = await extractAnswer(tab, query);
 		if (!answer) throw new Error("No answer captured from Gemini clipboard");
 
 		const finalUrl = await cdp(["eval", tab, "document.location.href"]).catch(

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@apmantza/greedysearch-pi",
-	"version": "1.8.9",
+	"version": "1.8.10",
 	"description": "Headless multi-engine AI search (Perplexity, Bing Copilot, Google AI) via browser automation -- NO API KEYS needed. Extracts answers with sources, optional synthesis. Grounded AI answers from real browser interactions.",
 	"type": "module",
 	"keywords": [

package/src/search/constants.mjs

@@ -1,44 +1,43 @@
-// src/search/constants.mjs — Shared constants for GreedySearch search pipeline
-
-import { tmpdir } from "node:os";
-
-export const GREEDY_PORT = 9222;
-export const GREEDY_PROFILE_DIR = `${tmpdir().replaceAll("\\", "/")}/greedysearch-chrome-profile`;
-export const ACTIVE_PORT_FILE = `${GREEDY_PROFILE_DIR}/DevToolsActivePort`;
-export const PAGES_CACHE = `${tmpdir().replaceAll("\\", "/")}/cdp-pages.json`;
-export const CHROME_MODE_FILE = `${tmpdir().replaceAll("\\", "/")}/greedysearch-chrome-mode`;
-
-// ALL_ENGINES drives the "all" fan-out. Add engines here to include them in multi-engine searches.
-// Engines in ENGINES but not in ALL_ENGINES are available for explicit use only.
-export const ALL_ENGINES = ["perplexity", "bing", "google"];
-
-export const ENGINE_DOMAINS = {
-	perplexity: "perplexity.ai",
-	bing: "copilot.microsoft.com",
-	google: "google.com",
-	gemini: "gemini.google.com",
-};
-
-export const ENGINES = {
-	perplexity: "perplexity.mjs",
-	pplx: "perplexity.mjs",
-	p: "perplexity.mjs",
-	bing: "bing-copilot.mjs",
-	bing2: "bing-aria.mjs",
-	copilot: "bing-copilot.mjs",
-	b: "bing-copilot.mjs",
-	google: "google-ai.mjs",
-	g: "google-ai.mjs",
-	gemini: "gemini.mjs",
-	gem: "gemini.mjs",
-	googlesearch: "google-search.mjs",
-	gs: "google-search.mjs",
-};
-
-export const SOURCE_FETCH_CONCURRENCY = Math.max(
-	1,
-	Number.parseInt(process.env.GREEDY_FETCH_CONCURRENCY || "4", 10) || 4,
-);
-
-// Tell cdp.mjs to prefer the GreedySearch Chrome profile's DevToolsActivePort
-process.env.CDP_PROFILE_DIR = GREEDY_PROFILE_DIR;
+// src/search/constants.mjs — Shared constants for GreedySearch search pipeline
+
+import { tmpdir } from "node:os";
+
+export const GREEDY_PORT = 9222;
+export const GREEDY_PROFILE_DIR = `${tmpdir().replaceAll("\\", "/")}/greedysearch-chrome-profile`;
+export const ACTIVE_PORT_FILE = `${GREEDY_PROFILE_DIR}/DevToolsActivePort`;
+export const PAGES_CACHE = `${tmpdir().replaceAll("\\", "/")}/cdp-pages.json`;
+export const CHROME_MODE_FILE = `${tmpdir().replaceAll("\\", "/")}/greedysearch-chrome-mode`;
+
+// ALL_ENGINES drives the "all" fan-out. Add engines here to include them in multi-engine searches.
+// Engines in ENGINES but not in ALL_ENGINES are available for explicit use only.
+export const ALL_ENGINES = ["perplexity", "bing", "google"];
+
+export const ENGINE_DOMAINS = {
+	perplexity: "perplexity.ai",
+	bing: "copilot.microsoft.com",
+	google: "google.com",
+	gemini: "gemini.google.com",
+};
+
+export const ENGINES = {
+	perplexity: "perplexity.mjs",
+	pplx: "perplexity.mjs",
+	p: "perplexity.mjs",
+	bing: "bing-copilot.mjs",
+	copilot: "bing-copilot.mjs",
+	b: "bing-copilot.mjs",
+	google: "google-ai.mjs",
+	g: "google-ai.mjs",
+	gemini: "gemini.mjs",
+	gem: "gemini.mjs",
+	googlesearch: "google-search.mjs",
+	gs: "google-search.mjs",
+};
+
+export const SOURCE_FETCH_CONCURRENCY = Math.max(
+	1,
+	Number.parseInt(process.env.GREEDY_FETCH_CONCURRENCY || "4", 10) || 4,
+);
+
+// Tell cdp.mjs to prefer the GreedySearch Chrome profile's DevToolsActivePort
+process.env.CDP_PROFILE_DIR = GREEDY_PROFILE_DIR;

package/extractors/bing-aria.mjs

@@ -1,539 +0,0 @@
-#!/usr/bin/env node
-
-// extractors/bing-aria.mjs — ARIA-tree-based Bing Copilot extractor
-//
-// Instead of copy button → clipboard polling → DOM fallback → iframe spelunking,
-// this extractor builds an ARIA accessibility tree of the page, finds the
-// Copilot answer region, and extracts structured text + sources directly.
-//
-// Inspiration: browser-use-rs extract_dom.js (Playwright's ariaSnapshot)
-//
-// Usage:
-//   node extractors/bing-aria.mjs "<query>" [--tab <prefix>]
-//
-// Output (stdout): JSON { answer, sources, query, url }
-// Errors to stderr only — stdout is always clean JSON for piping.
-
-import {
-	cdp,
-	formatAnswer,
-	getOrOpenTab,
-	handleError,
-	injectClipboardInterceptor,
-	jitter,
-	outputJson,
-	parseArgs,
-	parseSourcesFromMarkdown,
-	prepareArgs,
-	TIMING,
-	validateQuery,
-	waitForSelector,
-	waitForStreamComplete,
-} from "./common.mjs";
-import { dismissConsent, handleVerification } from "./consent.mjs";
-import { SELECTORS } from "./selectors.mjs";
-
-const S = SELECTORS.bing;
-
-// ============================================================================
-// ARIA-tree answer extraction
-// ============================================================================
-
-const EXTRACT_ARIA_JS = String.raw`
-(async function() {
-	'use strict';
-
-	// ── visibility helpers ──────────────────────────
-	function isHidden(el) {
-		if (['STYLE','SCRIPT','NOSCRIPT','TEMPLATE'].includes(el.tagName)) return true;
-		const s = window.getComputedStyle(el);
-		if (s.visibility !== 'visible') return true;
-		if (s.display === 'none') return true;
-		if (el.getAttribute('aria-hidden') === 'true') return true;
-		return false;
-	}
-
-	function isVisible(el) {
-		const r = el.getBoundingClientRect();
-		return r.width > 0 && r.height > 0;
-	}
-
-	function getRole(el) {
-		const explicit = el.getAttribute('role');
-		if (explicit) return explicit.split(' ')[0];
-		const tag = el.tagName;
-		const map = {
-			BUTTON:'button', A: el.hasAttribute('href')?'link':null,
-			INPUT: (()=>{ const t=(el.type||'text').toLowerCase(); return {button:'button',checkbox:'checkbox',radio:'radio',range:'slider',search:'searchbox',text:'textbox',email:'textbox',tel:'textbox',url:'textbox',number:'spinbutton'}[t]||'textbox'; })(),
-			TEXTAREA:'textbox', SELECT: el.hasAttribute('multiple')||el.size>1?'listbox':'combobox',
-			H1:'heading',H2:'heading',H3:'heading',H4:'heading',H5:'heading',H6:'heading',
-			IMG: el.getAttribute('alt')===''?'presentation':'img',
-			NAV:'navigation', MAIN:'main', ARTICLE:'article',
-			HEADER:'banner', FOOTER:'contentinfo', ASIDE:'complementary',
-			FORM:'form', TABLE:'table', UL:'list', OL:'list', LI:'listitem',
-			P:'paragraph', DIALOG:'dialog', IFRAME:'iframe'
-		};
-		return map[tag] || 'generic';
-	}
-
-	function getName(el) {
-		const label = el.getAttribute('aria-label');
-		if (label) return label;
-		const labelledBy = el.getAttribute('aria-labelledby');
-		if (labelledBy) {
-			const texts = labelledBy.split(/\s+/).map(id => {
-				const e = document.getElementById(id);
-				return e ? e.textContent : '';
-			}).filter(Boolean);
-			if (texts.length) return texts.join(' ');
-		}
-		if (['INPUT','TEXTAREA','SELECT'].includes(el.tagName)) {
-			const id = el.id;
-			if (id) {
-				const lbl = document.querySelector('label[for="'+id+'"]');
-				if (lbl) return lbl.textContent || '';
-			}
-			const parentLbl = el.closest('label');
-			if (parentLbl) return parentLbl.textContent || '';
-			const ph = el.getAttribute('placeholder');
-			if (ph) return ph;
-		}
-		if (el.tagName === 'IMG') return el.getAttribute('alt') || '';
-		const title = el.getAttribute('title');
-		if (title) return title;
-		return '';
-	}
-
-	// ── Build ARIA tree ──────────────────────────
-	let indexCounter = 0;
-
-	function buildTree(node, visited = new Set()) {
-		if (visited.has(node)) return null;
-		visited.add(node);
-
-		if (node.nodeType === 3) { // text node
-			return node.nodeValue;
-		}
-		if (node.nodeType !== 1) return null;
-
-		const el = node;
-		if (isHidden(el)) return null;
-
-		const role = getRole(el);
-		if (!role || role === 'presentation' || role === 'none') return null;
-
-		const name = (getName(el) || '').replace(/\s+/g, ' ').trim();
-		const box = el.getBoundingClientRect();
-		const visible = isVisible(el);
-		const cursor = window.getComputedStyle(el).cursor;
-
-		const result = {
-			role,
-			name,
-			children: [],
-			visible,
-			cursor,
-			tag: el.tagName,
-		};
-
-		// index visible interactive + pointer-cursor elements
-		if (visible && (cursor === 'pointer' || ['button','link','textbox','searchbox',
-			'checkbox','radio','combobox','listbox','option','menuitem',
-			'slider','spinbutton','switch','tab','heading'].includes(role))) {
-			result.index = indexCounter++;
-			if (el.tagName === 'A' && el.href) result.href = el.href;
-			if (el.id) result.id = el.id;
-		}
-
-		// shadow DOM
-		if (el.shadowRoot) {
-			for (let c = el.shadowRoot.firstChild; c; c = c.nextSibling) {
-				const child = buildTree(c, visited);
-				if (child) result.children.push(child);
-			}
-		}
-
-		// regular children
-		for (let c = el.firstChild; c; c = c.nextSibling) {
-			if (c.assignedSlot) continue;
-			const child = buildTree(c, visited);
-			if (child) result.children.push(child);
-		}
-
-		// aria-owns
-		if (el.hasAttribute('aria-owns')) {
-			for (const id of el.getAttribute('aria-owns').split(/\s+/)) {
-				const owned = document.getElementById(id);
-				if (owned) {
-					const child = buildTree(owned, visited);
-					if (child) result.children.push(child);
-				}
-			}
-		}
-
-		return result;
-	}
-
-	// ── Find answer region ──────────────────────────
-	// Locale-agnostic: finds the LAST Copilot AI message container.
-	// Copilot uses consistent CSS patterns across locales:
-	//   - AI messages: class contains "ai-message" or "response"
-	//   - User messages: different class prefix
-	//   - We take the LAST ai-message container in DOM order.
-	function findAnswerRegion() {
-		// Look for AI message containers with the known Copilot class pattern
-		// Tailwind-based: group/ai-message, .response-content, etc.
-		const allDivs = document.querySelectorAll('div[class*="ai-message"], div[class*="response-content"], div[class*="message"]');
-		let best = null;
-		for (const el of allDivs) {
-			const text = (el.innerText || '');
-			if (text.length > 100) best = el;  // take last one with substantial text
-		}
-		if (best) return best;
-
-		// Fallback: walk the DOM looking for containers with role=region/article
-		// and substantial text, take the last one.
-		const containers = [];
-		(function walk(el) {
-			if (!el || el.nodeType !== 1) return;
-			const role = el.getAttribute('role');
-			const cls = (el.className || '').toString();
-			if (role === 'region' || role === 'article' ||
-				cls.includes('ac-container')) {
-				const text = (el.innerText || '');
-				if (text.length > 100) containers.push(el);
-			}
-			for (const c of el.children) walk(c);
-		})(document.body);
-		if (containers.length > 0) return containers[containers.length - 1];
-
-		return document.body;
-	}
-
-	// ── Extract text from tree ──────────────────────
-	// Returns a normalized string. Block elements get newlines.
-	// Inline elements flow with whitespace separation.
-	// Locale-agnostic: filters by role/structure only, never by text content.
-	function extractText(node, isInline = false) {
-		if (typeof node === 'string') return node;
-		if (!node) return '';
-
-		// Skip UI buttons entirely
-		if (node.role === 'button') return '';
-
-		const parts = [];
-
-		// heading → markdown heading with surrounding newlines (block)
-		if (node.role === 'heading') {
-			const level = parseInt(node.tag?.[1]) || 2;
-			const inner = node.children.map(c => extractText(c)).join('');
-			if (inner.trim()) parts.push('\n' + '#'.repeat(level) + ' ' + inner.trim() + '\n');
-		}
-
-		// link → markdown link (inline if inside text, block otherwise)
-		else if (node.role === 'link' && node.href) {
-			const text = node.children.map(c => extractText(c)).join('').trim();
-			if (text && node.href.startsWith('http')) {
-				parts.push('[' + text + '](' + node.href + ')');
-			} else if (text) {
-				parts.push(text);
-			}
-		}
-
-		// listitem → preserve structure (block)
-		else if (node.role === 'listitem') {
-			const text = node.children.map(c => extractText(c)).join('').trim();
-			if (text) parts.push('\n- ' + text);
-		}
-
-		// code blocks
-		else if (node.tag === 'CODE' || node.tag === 'PRE') {
-			const text = node.children.map(c => extractText(c)).join('').trim();
-			if (text) parts.push('\n\x60\x60\x60\n' + text + '\n\x60\x60\x60\n');
-		}
-
-		// paragraph — block level, newlines around
-		else if (node.role === 'paragraph') {
-			const text = node.children.map(c => extractText(c)).join('').trim();
-			if (text) parts.push('\n' + text + '\n');
-		}
-
-		// generic/inline — flow text, join tight (whitespace already in text nodes)
-		else {
-			for (const child of node.children) {
-				parts.push(extractText(child));
-			}
-		}
-
-		return parts.join('');
-	}
-
-	// ── Collect sources ──────────────────────────
-	function collectLinks(node) {
-		const links = [];
-		function walk(n) {
-			if (typeof n === 'string') return;
-			if (!n) return;
-			if (n.role === 'link' && n.href && n.href.startsWith('http') &&
-				!n.href.includes('copilot.microsoft.com') &&
-				!n.href.includes('bing.com') &&
-				!n.href.includes('microsoft.com/privacy')) {
-				links.push({ title: n.name || '', url: n.href });
-			}
-			for (const c of n.children) walk(c);
-		}
-		walk(node);
-		// deduplicate by url
-		const seen = new Set();
-		return links.filter(l => { if (seen.has(l.url)) return false; seen.add(l.url); return true; });
-	}
-
-	// ── Execute ──────────────────────────────────
-	try {
-		// Wait for the answer to actually render — the stream may be "complete"
-		// but React hasn't painted the AI message yet. Poll for ai-message content.
-		await new Promise(r => setTimeout(r, 400));
-
-		const deadline = Date.now() + 8000;
-		let answerEl = null;
-		while (Date.now() < deadline) {
-			answerEl = findAnswerRegion();
-			if (answerEl && (answerEl.innerText || '').length > 200) break;
-			answerEl = null;
-			await new Promise(r => setTimeout(r, 500));
-		}
-
-		if (!answerEl) {
-			return JSON.stringify({ error: 'No answer region found (content too short or not rendered)', answer: '', sources: [] });
-		}
-
-		if (!answerEl) {
-			return JSON.stringify({ error: 'No answer region found', answer: '', sources: [] });
-		}
-
-		const tree = buildTree(answerEl);
-		if (!tree) {
-			return JSON.stringify({ error: 'ARIA tree build failed', answer: '', sources: [] });
-		}
-
-		const text = extractText(tree);
-
-		// Post-process: structural normalization only (locale-agnostic)
-		// Buttons are already filtered by role in extractText.
-		// Deduplication handles Copilot's mobile+desktop DOM variants.
-		let clean = text
-			.replace(/\n{3,}/g, '\n\n')
-			.replace(/^\s+|\s+$/g, '')
-			.trim();
-
-		// Strip leading heading if it's the Copilot "X said" label
-		// (locale-agnostic: just checks for markdown heading syntax at start)
-		clean = clean.replace(/^#{1,6}\s+.+?\n\n/, '');
-
-		// Deduplicate: Copilot sends duplicate DOM for responsive variants
-		const lines = clean.split('\n');
-		const seen = new Set();
-		const deduped = [];
-		for (const line of lines) {
-			const normalized = line.trim();
-			if (!normalized) { deduped.push(''); continue; }
-			if (normalized.length <= 2 && /^[-–—•·]$/.test(normalized)) continue;
-			if (seen.has(normalized)) continue;
-			seen.add(normalized);
-			deduped.push(line);
-		}
-		clean = deduped.join('\n').replace(/\n{3,}/g, '\n\n').trim();
-
-		const sources = collectLinks(tree).slice(0, 10);
-
-		return JSON.stringify({ answer: clean, sources });
-	} catch (e) {
-		return JSON.stringify({ error: e.toString(), answer: '', sources: [] });
-	}
-})()
-`;
-
-async function extractAnswer(tab) {
-	console.error("[bing-aria] Extracting answer via ARIA tree...");
-
-	const resultRaw = await cdp(["eval", tab, EXTRACT_ARIA_JS], 45000);
-
-	let result;
-	try {
-		result = JSON.parse(resultRaw);
-	} catch {
-		throw new Error(
-			`ARIA extraction returned invalid JSON: ${resultRaw.slice(0, 200)}`,
-		);
-	}
-
-	if (result.error) {
-		throw new Error(`ARIA extraction failed: ${result.error}`);
-	}
-
-	const { answer, sources: ariaSources } = result;
-
-	if (!answer || answer.length < 10) {
-		throw new Error(
-			`ARIA extraction returned insufficient content (${answer?.length || 0} chars)`,
-		);
-	}
-
-	// Hybrid: click copy button for markdown sources only (answer already extracted via ARIA).
-	// At this point the copy button is guaranteed rendered — no hydration race, no retries.
-	const GLOBAL_VAR = "__bingAriaClipboard";
-	await injectClipboardInterceptor(tab, GLOBAL_VAR);
-	const clipSources = await grabClipboardSources(tab, GLOBAL_VAR);
-	console.error(`[bing-aria] Clipboard sources: ${clipSources.length}`);
-
-	// Merge: ARIA DOM sources + clipboard markdown sources
-	const allSources = [...ariaSources, ...clipSources]
-		.filter((v, i, arr) => arr.findIndex((x) => x.url === v.url) === i)
-		.slice(0, 10);
-
-	console.error(
-		`[bing-aria] Extracted ${answer.length} chars, ${allSources.length} sources`,
-	);
-	return { answer: answer.trim(), sources: allSources };
-}
-
-/**
- * Click the Bing copy button, grab clipboard markdown, extract just the [title](url) sources.
- * Lightweight — no retries, no hydration delay (button is already visible at this point).
- */
-async function grabClipboardSources(tab, globalVar) {
-	try {
-		// Click last copy button (the most recent AI message)
-		const copyBtn = S.copyButton;
-		await cdp([
-			"eval",
-			tab,
-			`(() => {
-				window.${globalVar} = '';
-				const buttons = document.querySelectorAll('${copyBtn}');
-				buttons[buttons.length - 1]?.click();
-			})()`,
-		]);
-
-		// Poll clipboard briefly (2s max — if it doesn't work, no big deal)
-		const deadline = Date.now() + 2000;
-		while (Date.now() < deadline) {
-			const text = await cdp(["eval", tab, `window.${globalVar} || ''`]).catch(
-				() => "",
-			);
-			if (text && text.length > 20) {
-				return parseSourcesFromMarkdown(text);
-			}
-			await new Promise((r) => setTimeout(r, 200));
-		}
-	} catch (e) {
-		console.error(`[bing-aria] Clipboard source grab failed: ${e.message}`);
-	}
-	return [];
-}
-
-// ============================================================================
-// Main
-// ============================================================================
-
-const USAGE =
-	'Usage: node extractors/bing-aria.mjs "<query>" [--tab <prefix>]\n';
-
-async function main() {
-	const args = await prepareArgs(process.argv.slice(2));
-	validateQuery(args, USAGE);
-
-	const { query, tabPrefix, short } = parseArgs(args);
-
-	try {
-		if (!tabPrefix) await cdp(["list"]);
-		const tab = await getOrOpenTab(tabPrefix);
-
-		const currentUrl = await cdp(["eval", tab, "document.location.href"]).catch(
-			() => "",
-		);
-		let onCopilot = false;
-		try {
-			const host = new URL(currentUrl).hostname.toLowerCase();
-			onCopilot =
-				host === "copilot.microsoft.com" ||
-				host.endsWith(".copilot.microsoft.com");
-		} catch {}
-
-		if (!onCopilot) {
-			await cdp(["nav", tab, "https://copilot.microsoft.com/"], 20000);
-			await new Promise((r) => setTimeout(r, 600));
-		}
-		await dismissConsent(tab, cdp);
-
-		const verifyResult = await handleVerification(tab, cdp, 10000);
-		if (verifyResult === "needs-human") {
-			throw new Error(
-				"Copilot verification required — please solve it manually in the browser window",
-			);
-		}
-
-		if (verifyResult === "clicked") {
-			await new Promise((r) => setTimeout(r, TIMING.afterVerify));
-			const currentUrl = await cdp([
-				"eval",
-				tab,
-				"document.location.href",
-			]).catch(() => "");
-			let onCopilot = false;
-			try {
-				const host = new URL(currentUrl).hostname.toLowerCase();
-				onCopilot =
-					host === "copilot.microsoft.com" ||
-					host.endsWith(".copilot.microsoft.com");
-			} catch {}
-			if (!onCopilot) {
-				await cdp(["nav", tab, "https://copilot.microsoft.com/"], 20000);
-				await new Promise((r) => setTimeout(r, 600));
-				await dismissConsent(tab, cdp);
-			}
-		}
-
-		const inputReady = await waitForSelector(tab, S.input, 15000, 500);
-		await new Promise((r) => setTimeout(r, jitter(300)));
-
-		if (!inputReady) {
-			throw new Error(
-				"Copilot input not found — verification may have failed or page is in unexpected state",
-			);
-		}
-
-		// NO clipboard interceptor needed — ARIA extraction reads the DOM directly
-		await cdp(["click", tab, S.input]);
-		await new Promise((r) => setTimeout(r, TIMING.postClick));
-		await cdp(["type", tab, query]);
-		await new Promise((r) => setTimeout(r, TIMING.postType));
-
-		await cdp([
-			"eval",
-			tab,
-			`document.querySelector('${S.input}')?.dispatchEvent(new KeyboardEvent('keydown',{key:'Enter',bubbles:true,keyCode:13})), 'ok'`,
-		]);
-
-		// Wait for Copilot's response to finish streaming
-		await waitForStreamComplete(tab, { timeout: 60000, minLength: 50 });
-
-		const { answer, sources } = await extractAnswer(tab);
-		if (!answer)
-			throw new Error("No answer extracted — Copilot may not have responded");
-
-		const finalUrl = await cdp(["eval", tab, "document.location.href"]).catch(
-			() => "",
-		);
-		outputJson({
-			query,
-			url: finalUrl,
-			answer: formatAnswer(answer, short),
-			sources,
-		});
-	} catch (e) {
-		handleError(e);
-	}
-}
-
-main();