AI Kansoku Lab Tracker
| 开发者 | aikansokulab |
|---|---|
| 更新时间 | 2026年7月6日 14:41 |
| PHP版本: | 7.4 及以上 |
| WordPress版本: | 7.0 |
| 版权: | GPL-2.0-or-later |
| 版权网址: | 版权信息 |
详情介绍:
AI Kansoku Lab Tracker detects visits from AI-related crawlers such as GPTBot, ChatGPT-User, ClaudeBot, and PerplexityBot using PHP server-side detection.
AI crawlers often do not execute JavaScript, so JavaScript-based analytics tools such as GA4 may not record these visits as normal pageviews. This plugin helps site owners observe AI crawler activity at the server level.
Detected crawler visits are stored in your WordPress database. Optional external tracking is available only when you explicitly enable it in the settings page.
Features
- 34 AI-related crawler patterns from 20+ companies – GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, Anthropic-AI, PerplexityBot, Perplexity-User, Google-Extended, Google-NotebookLM, Google-Agent, ByteSpider, Meta-ExternalAgent, Meta-ExternalFetcher, AppleBot, AppleBot-Extended, DeepSeek, Grok, xAI-SearchBot, CopilotBot, Bingbot, AmazonBot, YandexBot, YouBot, Diffbot, ShapBot, TavilyBot, DuckAssistBot, Cohere-AI, Timpibot, FirecrawlAgent, and related access patterns
- Training vs. user-triggered classification – Crawler visits are classified as either "Autonomous AI (training/patrol)" or "Human-triggered AI reference," showing whether your content is being collected for AI training or actually being looked up because someone asked an AI about it
- Local crawler logging – Detected visits are stored in your WordPress database
- 30-day trend analysis – Dashboard with 30-day / 7-day period switching
- AI Exposure Score – Visibility score based on crawler activity
- Crawler breakdown – See which AI-related services visit your site most
- Top pages – Identify which pages AI crawlers focus on
- Explicit opt-in for external tracking – External transmission only activates when you enable it via checkbox
- Minimal performance impact – Optional external requests use non-blocking POST requests
- Front-end only – Does not fire on admin, cron, AJAX, or REST API requests
- WordPress standards – Uses wp_remote_post, register_setting, nonce verification, and proper escaping throughout
- 20 社以上・ 34 種類の AI 関連クローラーパターンを検知
- 「学習・巡回 AI」と「人間起点の AI 参照」を区別して表示 — サイトが AI の学習データとして収集されているだけなのか、それとも誰かが実際に AI に質問した結果として読まれているのかが一目でわかります
- 検知ログを WordPress のローカル DB に保存
- 30 日間のトレンド分析（7 日間表示への切り替え対応）
- AI 露出スコアによる可視化
- クローラー種別の内訳表示
- よく見られているページの可視化
- 外部トラッキングは明示的なオプトイン方式
- 管理画面・ AJAX ・ REST API ・ WP-Cron では実行しない設計
- User-Agent string of the detected crawler
- URL path of the visited page
- HTTP referer header, if present
- HTTP status code of the page, such as 200 or 404
- Site ID configured in the plugin settings
- WordPress 6.0 or higher
- PHP 7.4 or higher
安装:
- Upload the plugin ZIP via Plugins > Add New > Upload Plugin
- Activate the plugin
- Go to Settings > AI Kansoku Lab
- Review the local crawler tracking dashboard
- Optional: Enter your Site ID from https://www.ai-kansoku.com/dashboard
- Optional: Check "Enable external tracking to AI Kansoku Lab"
- Click Save Settings
常见问题:
Where do I find my Site ID?
Log in to AI Kansoku Lab at https://www.ai-kansoku.com, register your site, and copy the Site ID shown in your dashboard. A Site ID is only required if you want to enable external tracking and use the AI Kansoku Lab external dashboard.
Will this slow down my site?
The plugin performs a lightweight User-Agent check on front-end requests.
When external tracking is enabled, the plugin uses wp_remote_post with blocking set to false, so the optional HTTP request is fired asynchronously and does not delay page rendering.
Does it run on every request?
No. The plugin is designed for front-end page requests only. Admin pages, AJAX calls, REST API requests, and WP-Cron jobs are excluded. External tracking only runs when both a Site ID is set and external tracking is explicitly enabled.
Can it detect crawlers that do not run JavaScript?
Yes. That is the main purpose of this plugin. JavaScript-based analytics tools may miss crawlers that do not execute JavaScript. This plugin detects them at the PHP level using the HTTP_USER_AGENT server variable.
Which AI-related crawlers are supported?
The plugin currently detects 34 crawler patterns from 20+ companies, including GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, Anthropic-AI, PerplexityBot, Perplexity-User, Google-Extended, Google-NotebookLM, Google-Agent, Googlebot, ByteSpider, Meta-ExternalAgent, Meta-ExternalFetcher, AppleBot, AppleBot-Extended, DeepSeek, Grok, xAI-SearchBot, CopilotBot, Bingbot, AmazonBot, YandexBot, YouBot, Diffbot, ShapBot, TavilyBot, DuckAssistBot, Cohere-AI, Timpibot, and FirecrawlAgent. Because new AI crawlers appear frequently, this list may lag slightly behind the live detection patterns shipped in the plugin. The full, up-to-date list is always visible in the "Crawler Breakdown" section of your dashboard.
What does "Autonomous AI" vs. "Human-triggered AI reference" mean?
Starting in version 1.4.11, detected visits are classified into two groups:
- Autonomous AI (training/patrol) – Crawlers like GPTBot or ClaudeBot that periodically and automatically collect content, typically for AI model training or search indexing.
- Human-triggered AI reference – Crawlers like ChatGPT-User, Claude-User, Perplexity-User, or Google-NotebookLM that only fetch a page because a person asked an AI assistant about that specific URL.
What data is stored locally?
Detected crawler visits are stored in your WordPress database. The stored data is used to show crawler trends, crawler breakdowns, top pages, and related dashboard information.
Is external tracking required?
No. Local crawler logging works without enabling external tracking. External tracking is optional and only activates when you explicitly enable it in the plugin settings page.
What happens when I uninstall the plugin?
All plugin settings stored in the WordPress database are removed automatically on uninstall.
更新日志:
1.4.13
- Improved: Replaced inline
<script>blocks in AI Crawled Pages, AI Visit Log, and Improvement Actions screens with properly enqueued external JavaScript files (wp_enqueue_script/wp_localize_script), following WordPress.org coding guidelines - Fixed: Pagination buttons on the AI Crawled Pages screen no longer overflow off-screen when there are many pages (added flex-wrap)
- Improved: AJAX handlers for visit memos and improvement actions now apply
wp_unslash()consistently to POST/GET input - Improved: Improvement Actions AJAX handlers now validate
action_typeagainst an allowlist and return proper HTTP status codes on error - Improved: Visit memo save now verifies the target log entry exists before writing, and the memo table uses a UNIQUE index on log_id
- Fixed: Status dropdown on the Improvement Actions screen now correctly rolls back to the previous value if a status update fails, using a dedicated
data-current-statusattribute - Improved: JavaScript-side error handling now reads structured error messages (
{ message: ... }) from the server and falls back gracefully; added network-failure handling (.catch) to all AJAX calls
- 改善: AI 訪問ログ画面の高速化。同一 URL への累積訪問回数を表示するための重い集計処理を廃止
- 改善: 本日の訪問数・ Bot 種類数などの集計を 5 分間キャッシュし、管理画面の表示速度を向上
- 改善: ログ一覧の絞り込み・並び替え用インデックスを追加（crawler, detected_at）
- 改善: AI 訪問ログのバッジ表示を「人間起点の AI 参照」「定期観測（自律巡回）」の分類に統一（ダッシュボードと同じ基準に）
- 改善: DB スキーマ更新処理を 1 回のみ実行する方式に変更し、通常表示時の負荷を軽減
- 修正: 日次集計の日付判定ロジックを見直し、インデックスが正しく効くように修正
- Added: Crawler visits are now classified as "Autonomous AI (training/patrol)" or "Human-triggered AI reference"
- Added: Two new first-view dashboard cards showing the autonomous vs. human-triggered breakdown
- Changed: "Top Page" and "AI Reach Rate" cards moved from the dashboard first view to the detailed analytics section below
- Added: Claude-User crawler pattern (Anthropic)
- Improved: Crawler detection now resolves both legacy name-based and current key-based log entries when determining classification
- セキュリティ改善: SaaS 連携の認証を X-AIKLab-Secret ヘッダー方式に変更
- 旧方式との一時的な互換性を維持
- Refactored AI bot detection definitions into a centralized bot master.
- Added and adjusted bot definitions for xAI SearchBot, Meta External Fetcher, and related crawlers.
- Added: 30-day dashboard view as default, previously 7-day only
- Added: Period switching UI with 30-day / 7-day toggle on dashboard
- Improved: KPI cards, charts, and crawler breakdown now respond to period selection
- Improved: Score calculation normalized by daily average for consistent scoring across periods
- Improved: Status panel shows accurate state when no data is available
- Fixed: Insight spike detection now uses actual 7-day data regardless of selected period
- Fixed: secret_key was not saved correctly in some environments
- Fixed: Settings page CSS/JS loading issue in some environments
- Fixed: Local crawler logging now works without external service connection
- Improved: REST API authentication moved to permission_callback with site_id and secret_key verification
- Improved: External API transmission only occurs when site_id is configured
- Removed all PRO/free feature restrictions
- Removed Stripe payment integration
- Removed AI citation checker
- Replaced Chart.js CDN with locally bundled file
- Removed inline script/style output
- Improved output escaping throughout
- Updated readme to disclose transmitted data fields including referer and status_code
- Renamed plugin to Kansoku Lab Tracker for WordPress.org guidelines compliance
- Updated all prefixes from AIK_ to AIKLAB_
- Replaced inline style block with proper enqueued styles
- Updated Terms of Service URL
- Replaced HTML-mixed echo with PHP template syntax in status panel
- Changed Tested up to from 6.9 to 6.8 for accuracy
- Changed hook from init to template_redirect for cleaner front-end-only targeting
- Removed redundant admin, cron, AJAX, and REST guards
- Replaced Markdown link syntax in readme with plain URLs for better compatibility
- Added explicit opt-in checkbox for external tracking
- Unified all URLs to https://www.ai-kansoku.com
- Refined crawler regex
- Updated privacy wording to accurately reflect transmitted data
- Status panel now shows tracking state
- Changed tracking from GET to POST request
- Increased timeout from 0.1s to 1s for better reliability
- Restricted tracking to front-end requests only
- Added explicit External API Notice in settings page
- Added uninstall.php to clean up database options on removal
- Initial release