內容簡介
Forge AI Bot Log 外掛可幫助網站管理者辨識和管理訪問網站的 AI 爬蟲，透過驗證技術確保流量的真實性，並提供詳細的爬蟲活動報告，讓使用者能夠有效控制 AI 流量的影響。
【主要功能】
• 實時儀表板：顯示過去 24 小時、7 天及 30 天的爬蟲訪問數據
• 爬蟲目錄：提供各爬蟲的擁有者及功能說明
• 驗證標籤：對每次訪問進行驗證，標記為已驗證、偽裝或未驗證
• 訪問控制：一鍵封鎖任何爬蟲，並更新 robots.txt
• 每月電子郵件摘要：每月 1 日發送上月活動報告
• CSV 匯出：下載完整或過濾的活動日誌進行分析
外掛標籤
開發者團隊
② 後台搜尋「Forge AI Bot Log – Block AI Bots, Crawler Detection & Verification」→ 直接安裝（推薦）
原文外掛簡介
Any script can claim to be GPTBot. This plugin checks whether it really is.
Most AI bot trackers read the user-agent string and take it at face value — so a scraper that sends User-agent: GPTBot gets logged as OpenAI. Forge AI Bot Log verifies each visit against the crawler’s own published infrastructure using forward-confirmed reverse DNS — the technique Google, OpenAI and Anthropic recommend for confirming their own crawlers — and labels every hit Verified, Spoofed or Unverified. You find out which AI traffic genuinely came from OpenAI, and which was something else wearing its name.
Verification covers every crawler that publishes a way to check it, including OpenAI, Anthropic, Google, Perplexity, Amazon, Apple and Meta. Bots with no published verification method are labelled Unverified rather than guessed at.
Beyond verification, Forge AI Bot Log shows you exactly which AI crawlers and assistants are visiting your WordPress site and explains what each one actually does. The bundled directory covers 23 bots from 15 companies, including GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-User, Google-Extended, GoogleOther, PerplexityBot, CCBot, Bytespider, Amazonbot, Applebot, Applebot-Extended, meta-externalagent, cohere-ai, Diffbot, YouBot, MistralAI-User and DuckAssistBot.
Key features
Live dashboard — bot visits for the last 24 hours, 7 days, 30 days, plus a complete searchable log.
Bot directory — click any bot for a plain-English explanation of who owns it, whether it trains AI models on your pages, powers AI search, or fetches pages live, and whether it obeys robots.txt.
Most-crawled content — see which of your pages AI bots are reading most.
Verified / Spoofed labelling — forward-confirmed reverse DNS on every tracked hit, so impersonation is flagged instead of trusted. Results are cached and throttled to keep it off the critical path.
Access control — block any bot with one click. Blocking adds a Disallow: / directive for that crawler to your robots.txt, and there is a hard-block .htaccess snippet for the bots that ignore robots.txt.
Monthly email summary — sent to the site admin on the 1st of each month at 9am, covering the previous month.
New-bot alerts — optional email the first time a brand-new AI bot appears.
Reliable capture — an in-PHP logger, an optional must-use early logger, and a server access-log importer for sites behind full-page caching or a CDN.
Privacy-friendly — only logs bots, not human visitors, with optional IP anonymisation and configurable data retention.
CSV export — download the full or filtered activity log for your own analysis.
Why it matters
AI crawlers may use your content as training data, cite you in AI search answers, or read your pages on a user’s behalf. Which of those you want is your call — but you cannot make that call without seeing the traffic first. Forge AI Bot Log makes AI bot activity visible and gives you the controls to manage it.
Who this is for
Publishers and bloggers deciding whether to opt out of LLM training. Site owners who suspect content scraping and want evidence. Agencies monitoring client sites, who need to monitor crawler activity across a portfolio. Anyone who has read that they should block AI bots but wants to know which bots are actually visiting before blocking anything. If you want an AI bot blocker that shows you its work rather than a blocklist you have to trust, this is it.
Google-Extended and Applebot-Extended are opt-out tokens rather than crawlers in their own right, and the plugin treats them as such — so you can decline LLM training without touching normal search crawling.
External services
Cross-promotion feed (optional). If you enter a promo feed URL in Settings, the plugin fetches that URL to populate the “more plugins” block in its own emails. Only the request is made; no site, user, or visitor data is transmitted. No feed URL is configured by default.
