extractor.js

// =============================================================================
// Chat Bird's-Eye — Universal Chat Extractor (bookmarklet source)
//
// Click this bookmarklet on any AI chat page. It detects the site, scrapes the
// currently-visible conversation(s), accumulates them in localStorage, and
// offers a one-click export to birdseye_chats.json — ready to load into the
// Chat Bird's-Eye workspace.
//
// Supported sites (auto-detected by hostname):
//   chatgpt.com / chat.openai.com   →  ChatGPT
//   claude.ai                        →  Claude
//   gemini.google.com                →  Gemini
//   perplexity.ai                    →  Perplexity
//   x.com (grok)                     →  Grok
//   copilot.microsoft.com            →  Copilot
//   *                                →  Generic fallback (best-effort DOM scrape)
//
// This file is the readable source. The actual bookmarklet is the minified
// single-line version produced by build-bookmarklet.js and embedded in
// index.html's "Get real chats" modal.
// =============================================================================
(function () {
  "use strict";
  if (window.__BIRDS_EYE_OPEN) { return; } // prevent double-injection
  window.__BIRDS_EYE_OPEN = true;

  var STORE_KEY = "birdseye_chats_v1";
  var host = location.hostname.replace(/^www\./, "");
  var site =
    /chatgpt\.com|chat\.openai\.com/.test(host) ? "ChatGPT" :
    /claude\.ai/.test(host) ? "Claude" :
    /gemini\.google\.com/.test(host) ? "Gemini" :
    /perplexity\.ai/.test(host) ? "Perplexity" :
    /x\.com/.test(host) && /grok/i.test(location.pathname) ? "Grok" :
    /copilot\.microsoft\.com/.test(host) ? "Copilot" :
    "Unknown";

  // ---- helpers ----
  function txt(el) { return (el.innerText || el.textContent || "").trim(); }
  function nowISO() { return new Date().toISOString(); }
  function uid() { return "ext_" + Date.now().toString(36) + Math.random().toString(36).slice(2, 6); }

  function loadStore() {
    try { return JSON.parse(localStorage.getItem(STORE_KEY) || "[]"); }
    catch (e) { return []; }
  }
  function saveStore(arr) {
    try { localStorage.setItem(STORE_KEY, JSON.stringify(arr)); } catch (e) {}
  }

  // ---- per-site scrapers. Each returns { title, messages: [{role, content, ts}] } or null ----

  function scrapeChatGPT() {
    // visible conversation turns
    var turns = document.querySelectorAll('[data-testid^="conversation-turn-"]');
    if (!turns.length) {
      // fallback: older DOM
      turns = document.querySelectorAll('article[data-testid]');
    }
    var msgs = [];
    for (var i = 0; i < turns.length; i++) {
      var t = turns[i];
      var role = /-2\b|-4\b|user/i.test(t.getAttribute("data-testid") || "") ? "user" : "assistant";
      // try to find the actual text container
      var content = txt(t);
      if (content) msgs.push({ role: role, content: content, ts: nowISO() });
    }
    if (!msgs.length) return null;
    var title = document.title.replace(/^ChatGPT\s*[:\-]?\s*/i, "").trim() || "Untitled ChatGPT";
    return { title: title, messages: msgs };
  }

  function scrapeClaude() {
    var userMsgs = document.querySelectorAll('[data-testid="user-message"]');
    var allMsgs = [];
    // Claude interleaves user + assistant in the main thread
    var thread = document.querySelector('[class*="conversation"]') || document.querySelector("main");
    if (!thread) return null;
    var blocks = thread.querySelectorAll('[data-testid="user-message"], [class*="assistant"]');
    if (blocks.length) {
      for (var i = 0; i < blocks.length; i++) {
        var b = blocks[i];
        var role = b.getAttribute("data-testid") === "user-message" ? "user" : "assistant";
        var content = txt(b);
        if (content) allMsgs.push({ role: role, content: content, ts: nowISO() });
      }
    }
    // fallback: just user messages
    if (!allMsgs.length && userMsgs.length) {
      for (var j = 0; j < userMsgs.length; j++) {
        var c = txt(userMsgs[j]);
        if (c) allMsgs.push({ role: "user", content: c, ts: nowISO() });
      }
    }
    if (!allMsgs.length) return null;
    var title = document.title.replace(/^Claude\s*[:\-]?\s*/i, "").trim() || "Untitled Claude";
    return { title: title, messages: allMsgs };
  }

  function scrapeGemini() {
    // Gemini renders turns in a scrollable conversation; try broad selectors
    var turns = document.querySelectorAll("message-content, .response-container, .model-response-text, .user-query, .query-text");
    var msgs = [];
    for (var i = 0; i < turns.length; i++) {
      var el = turns[i];
      var cls = (el.className || "") + " " + (el.tagName || "");
      var role = /user|query/i.test(cls) ? "user" : "assistant";
      var content = txt(el);
      if (content) msgs.push({ role: role, content: content, ts: nowISO() });
    }
    if (!msgs.length) return null;
    var title = document.title.replace(/^Gemini\s*[:\-]?\s*/i, "").trim() || "Untitled Gemini";
    return { title: title, messages: msgs };
  }

  function scrapePerplexity() {
    var queries = document.querySelectorAll('[class*="query"], textarea');
    var responses = document.querySelectorAll('[class*="prose"], [class*="answer"]');
    var msgs = [];
    // approximate: interleave by DOM order
    var all = document.querySelectorAll('[class*="query"], [class*="prose"]');
    for (var i = 0; i < all.length; i++) {
      var el = all[i];
      var role = /query/i.test(el.className || "") ? "user" : "assistant";
      var content = txt(el);
      if (content) msgs.push({ role: role, content: content, ts: nowISO() });
    }
    if (!msgs.length) return null;
    var title = document.title.replace(/^Perplexity\s*[:\-]?\s*/i, "").trim() || "Untitled Perplexity";
    return { title: title, messages: msgs };
  }

  function scrapeGeneric() {
    // best-effort: find the largest text container and split by headings/paragraphs
    var main = document.querySelector("main") || document.body;
    var paras = main.querySelectorAll("p, pre, h1, h2, h3");
    var msgs = [];
    var buf = "";
    for (var i = 0; i < paras.length; i++) {
      var t = txt(paras[i]);
      if (t.length > 10) buf += t + "\n\n";
    }
    if (buf.length > 50) msgs.push({ role: "user", content: buf.trim(), ts: nowISO() });
    if (!msgs.length) return null;
    return { title: document.title.trim() || "Untitled", messages: msgs };
  }

  var scrapers = {
    "ChatGPT": scrapeChatGPT,
    "Claude": scrapeClaude,
    "Gemini": scrapeGemini,
    "Perplexity": scrapePerplexity,
    "Grok": scrapeGeneric,
    "Copilot": scrapeGeneric,
    "Unknown": scrapeGeneric
  };

  var result = (scrapers[site] || scrapeGeneric)();
  if (!result) {
    toast("Couldn't find a chat on this page (" + site + "). Open a conversation and try again.", "#f87171");
    window.__BIRDS_EYE_OPEN = false;
    return;
  }

  // dedupe by title+first-message hash
  var store = loadStore();
  var dedupeKey = result.title + "|" + (result.messages[0] || {}).content.slice(0, 80);
  var exists = store.some(function (c) { return c._dedupe === dedupeKey; });
  if (!exists) {
    store.push({
      id: uid(),
      source: site,
      title: result.title,
      createdAt: nowISO(),
      updatedAt: nowISO(),
      messages: result.messages,
      _dedupe: dedupeKey
    });
    saveStore(store);
  }

  var count = store.length;
  toast(
    (exists ? "Already had that one. " : "Captured from " + site + ". ") +
    "Total: " + count + " chat" + (count === 1 ? "" : "s") + " stored. " +
    "Click here to export →",
    exists ? "#fbbf24" : "#4ade80",
    function () { exportJSON(store); }
  );

  // ---- toast UI ----
  function toast(msg, color, onClick) {
    var old = document.getElementById("__birdseye_toast");
    if (old) old.remove();
    var t = document.createElement("div");
    t.id = "__birdseye_toast";
    t.style.cssText = [
      "position:fixed", "bottom:20px", "right:20px", "z-index:2147483647",
      "background:#1a1a24", "color:#e8e8f0", "padding:14px 18px",
      "border-radius:12px", "font:14px/1.5 -apple-system,sans-serif",
      "max-width:380px", "box-shadow:0 8px 30px rgba(0,0,0,.5)",
      "border:1px solid " + (color || "#7c83ff"), "cursor:pointer",
      "transition:opacity .3s"
    ].join(";") + ";";
    t.textContent = msg;
    t.onclick = function () { if (onClick) onClick(); };
    document.body.appendChild(t);
    setTimeout(function () { t.style.opacity = "0"; setTimeout(function () { t.remove(); }, 400); }, 8000);
  }

  function exportJSON(store) {
    // strip internal _dedupe keys
    var clean = store.map(function (c) {
      return { id: c.id, source: c.source, title: c.title, createdAt: c.createdAt, updatedAt: c.updatedAt, messages: c.messages };
    });
    var blob = new Blob([JSON.stringify(clean, null, 2)], { type: "application/json" });
    var url = URL.createObjectURL(blob);
    var a = document.createElement("a");
    a.href = url; a.download = "birdseye_chats.json";
    document.body.appendChild(a); a.click(); a.remove();
    URL.revokeObjectURL(url);
    // clear store after export so user can start fresh
    if (confirm("Exported " + clean.length + " chats. Clear the bookmarklet's stored chats?")) {
      localStorage.removeItem(STORE_KEY);
    }
  }

  window.__BIRDS_EYE_OPEN = false;
})();