Bookmarklets

A bookmarklet is a bookmark whose address is a small JavaScript program instead of a URL. Clicking it runs that program against the page you are currently looking at — no extension to install, no permissions to grant, nothing that keeps running in the background.

How to install one

1. Make sure your bookmarks bar is visible (Ctrl+Shift+B, or Cmd+Shift+B on a Mac).
2. Drag the blue button below onto that bar.
3. Open the conversation you want and click the new bookmark.

If dragging is awkward, use Copy and paste the result into the URL field of a manually created bookmark.

• Export Claude Conversation

  Downloads the claude.ai conversation you are viewing as a Markdown file, including thinking blocks, tool calls and attachments.

  Export Claude Conversation Copy

  Source — claude-export.js

  /**
   * Claude.ai conversation → Markdown export bookmarklet
   *
   * Run it while viewing https://claude.ai/chat/<uuid>.
   * It uses Claude's own internal JSON API (same session cookies the page
   * uses), so it captures the full conversation without DOM scraping —
   * including messages scrolled out of view / virtualized away.
   *
   * Output: a downloaded .md file. Filename defaults to
   *   claude_<slugified-title>_<short-id>.md
   * and a prompt() lets you edit it before saving.
   *
   * @title Export Claude Conversation
   * @description Downloads the claude.ai conversation you are viewing as a Markdown
   *              file, including thinking blocks, tool calls and attachments.
   * @order 10
   */
  (async () => {
    try {
      // 1. Conversation id from the URL
      const m = location.pathname.match(/\/chat\/([0-9a-f-]{36})/i);
      if (!m) {
        alert('Open a conversation first (URL must look like /chat/<uuid>).');
        return;
      }
      const convId = m[1];
  
      // 2. Organization id: prefer the lastActiveOrg cookie, fall back to API
      let orgId = (document.cookie.match(/(?:^|;\s*)lastActiveOrg=([^;]+)/) || [])[1];
      if (!orgId) {
        const orgs = await (await fetch('/api/organizations', { credentials: 'include' })).json();
        const org = orgs.find(o => (o.capabilities || []).includes('chat')) || orgs[0];
        orgId = org && org.uuid;
      }
      if (!orgId) { alert('Could not determine organization id — are you logged in?'); return; }
  
      // 3. Fetch the full conversation tree
      const url = `/api/organizations/${orgId}/chat_conversations/${convId}` +
                  `?tree=True&rendering_mode=messages&render_all_tools=true`;
      const res = await fetch(url, { credentials: 'include' });
      if (!res.ok) { alert(`API request failed (${res.status}). Claude may have changed its API.`); return; }
      const conv = await res.json();
  
      // 4. Render one message's content blocks as markdown
      const renderBlock = (b) => {
        switch (b.type) {
          case 'text':
            return b.text || '';
          case 'thinking':
            return b.thinking
              ? `<details><summary>Thinking</summary>\n\n${b.thinking}\n\n</details>`
              : '';
          case 'tool_use':
            return '```json\n' +
                   `// tool_use: ${b.name || 'unknown'}\n` +
                   JSON.stringify(b.input || {}, null, 2) + '\n```';
          case 'tool_result': {
            const txt = Array.isArray(b.content)
              ? b.content.map(c => c.text || '').join('\n')
              : (typeof b.content === 'string' ? b.content : JSON.stringify(b.content));
            return '```\n' + `[tool_result${b.is_error ? ' (error)' : ''}]\n` + txt + '\n```';
          }
          default:
            return b.text || '';
        }
      };
  
      const renderMsg = (msg) => {
        const role = msg.sender === 'human' ? 'Human' : 'Assistant';
        const parts = (msg.content && msg.content.length)
          ? msg.content.map(renderBlock).filter(Boolean)
          : [msg.text || ''];
        // Attachments / uploaded files, if any
        const extras = [];
        (msg.attachments || []).forEach(a => {
          extras.push(`> **Attachment:** ${a.file_name || 'file'}` +
            (a.extracted_content ? `\n\n\`\`\`\n${a.extracted_content}\n\`\`\`` : ''));
        });
        (msg.files || []).forEach(f => extras.push(`> **File:** ${f.file_name || f.file_kind || 'file'}`));
        return `## ${role}\n\n${[...extras, ...parts].join('\n\n')}`;
      };
  
      const msgs = (conv.chat_messages || [])
        .slice()
        .sort((a, b) => new Date(a.created_at) - new Date(b.created_at));
  
      const title = conv.name || 'Untitled conversation';
      const md = [
        `# ${title}`,
        '',
        `- **Conversation ID:** ${convId}`,
        `- **URL:** ${location.origin}/chat/${convId}`,
        `- **Created:** ${conv.created_at || 'unknown'}`,
        `- **Exported:** ${new Date().toISOString()}`,
        `- **Messages:** ${msgs.length}`,
        '',
        '---',
        '',
        msgs.map(renderMsg).join('\n\n---\n\n'),
        ''
      ].join('\n');
  
      // 5. Filename: derived from title + short id, editable via prompt()
      const slug = title.toLowerCase().replace(/[^a-z0-9]+/g, '-').replace(/^-+|-+$/g, '').slice(0, 60);
      const suggested = `claude_${slug || 'conversation'}_${convId.slice(0, 8)}.md`;
      const filename = prompt('Save conversation as:', suggested);
      if (!filename) return; // user cancelled
  
      // 6. Download
      const blob = new Blob([md], { type: 'text/markdown;charset=utf-8' });
      const a = document.createElement('a');
      a.href = URL.createObjectURL(blob);
      a.download = filename.endsWith('.md') ? filename : filename + '.md';
      document.body.appendChild(a);
      a.click();
      setTimeout(() => { URL.revokeObjectURL(a.href); a.remove(); }, 1000);
    } catch (e) {
      alert('Export failed: ' + e.message);
      console.error(e);
    }
  })();

  View on GitHub

Source and licence

Every bookmarklet above is generated from the cleartext file linked under its source listing. The short version you drag is produced at build time — it is never stored as a second copy, so what you read is exactly what you run.

Contributions and bug reports are welcome on the repository.