知识库

Eye(FR) → Mill(skill) → Library(/app)

← 返回列表
Hacker News · 2026-05-14 · 已成文 · 来源 file

Arrivl 用确定性 UA/IP/路径规则识别 Agent 流量(训练爬虫、AI 搜索索引、agentic browser),并连到 bot→human 漏斗与 AI Readiness 审计。

为什么重要

站点侧可观测性决定内容如何被 Agent 消费;对公开知识站/文档站的 MCP 与 SEO 工程有直接动作。

正文

Arrivl 用确定性 UA/IP/路径规则识别 Agent 流量(训练爬虫、AI 搜索索引、agentic browser),并连到 bot→human 漏斗与 AI Readiness 审计。

Claim

Arrivl 用确定性 UA/IP/路径规则识别 Agent 流量(训练爬虫、AI 搜索索引、agentic browser),并连到 bot→human 漏斗与 AI Readiness 审计。

Why it matters

站点侧可观测性决定内容如何被 Agent 消费;对公开知识站/文档站的 MCP 与 SEO 工程有直接动作。

Summary

Agent 流量分析产品:意图分类 + readiness 审计;文档站可先自建 robots/llms.txt 清单。

Actions

  • (none)

Evidence

  • (none)

Caveats

  • (none)

Research queries

  • (none)

Body

正文

背景

传统分析忽略 AI agent 访问;推荐与答案依赖 Agent 如何抓取站点。

机制

意图分类(Research/Indexing/Training/Agentic 等);dashboard 含 referred humans、bot→human CVR;14 项 AI Readiness;Intake 基于 UA 注册表与路径 allowlist。

取舍

确定性匹配可解释但会漏自定义 Agent;营销产品需验证数据质量。

动作

先修 robots/llms.txt 与自建日志分类,再考虑第三方面板。

结合的源文章

主源
Show HN: Arrivl – Analytics for AI agent traffic on your site
打开原文 ↗

原文快照

展开 / 收起快照

As LLMs becoming a growing channel, many business notice their web traffic has dropped a lot in the past year but there is no way to tell how much of it was LLMs eating the clicks.

It's mainly due to most analytic tools (like GA4, Mixpanel, Amplitude) reply on a JS snippet that fires in a browser. But AI agents don't load JS. They hit your server, read the HTML, and leave. The only place they show up is in raw server logs.

So we built Arrivl. It works off server logs to show you which agents visited, what pages they read and whether they bring you real human visit. It also runs an audit of how AI friendly your site is and tells you how to improve in an afternoon.

It's completely free with no credit card required. https://arrivl.ai Install is a snippet or a log forwarder, takes a few minutes.

Would love some feedbacks from the community here.


Comments URL: https://news.ycombinator.com/item?id=48126836

Points: 2

# Comments: 0