Skip to main content
The Traffic view reports Humans and Agents as separate counts, and traffic exports break views into five columns. This page explains the two decisions behind those numbers: when Mintlify serves a page as Markdown and how views populate each category.

​
When Mintlify serves Markdown

Every page is available as HTML and as Markdown. Mintlify’s edge picks one per request, before rendering. A request receives Markdown when any of the following is true:
  1. The path ends in .md. For example, /analytics/traffic.md always returns Markdown.
  2. The Accept header prefers Markdown. Mintlify orders the header’s media types by quality value. It then compares the first Markdown type (text/markdown, text/x-markdown, or text/plain) with the first HTML type (text/html or application/xhtml+xml). If a Markdown type is present and either no HTML type is present or Markdown ranks higher, Mintlify returns Markdown.
  3. The user agent is a known AI agent and does not ask for HTML. If the user agent contains one of the following identifiers and the Accept header does not rank HTML above Markdown, Mintlify returns Markdown even when the request never mentions Markdown.
The identifiers that trigger rule 3 are Claude-User, ChatGPT-User, OAI-SearchBot, PerplexityBot, Perplexity-User, Google-Agent, MistralAI-User, DuckAssistBot, and cohere-ai. Matching is a case-insensitive substring check. Negotiation applies only to documentation pages. Static assets, llms.txt, skill.md, RSS, PDF exports, MCP endpoints, login pages, and internal /_mintlify/ routes are never negotiated. Mintlify caches HTML and Markdown responses as separate variants keyed on the Accept and User-Agent headers, so a browser never receives a cached Markdown response.

​
How a request becomes a view

Mintlify records two kinds of page view. Each view carries the request path, referrer, IP address, and user agent. Mintlify stores timestamps in UTC.

​
How views are categorized

Classification uses only the user agent string. Mintlify checks it, case-insensitively, against four ordered lists of known bot and agent identifiers. The first list that matches sets the category. No match means human. See the full user agent list for every identifier in each category. Mintlify excludes its own traffic, such as cache revalidation and link checking, before counting.

​
How the dashboard counts humans and agents

The dashboard combines the view type with the category:
  • Humans are HTML page views in the human category.
  • Agents are every view in the AI answers category, plus Markdown views in the human category.
Markdown views count as agent traffic even when the user agent looks like a browser or a generic HTTP client. Requesting Markdown is how agents read documentation, so the request itself is the signal. These views appear in the agent breakdown under the Browser and HTTP clients buckets. Search index, training, and other AI views appear in neither count. They are visible only in exports.
Classification trusts the user agent a client sends. An agent that identifies itself with a standard browser user agent and fetches HTML counts as a human view. A person who opens a .md URL directly counts as an agent view.

​
Full user agent list

Mintlify checks the user agent against these identifiers in the order shown. Matching is a case-insensitive substring check, so chatgpt matches ChatGPT-User and any other user agent containing that text.

​
AI answers

ChatGPT-User, chatgpt, Claude-User, claude-code, Cursor, Opencode, WarpBot, Perplexity-User, Google-Agent, GoogleAgent, Google-NotebookLM, Google-Read-Aloud, MistralAI-User, Copilot, Amzn-User, DuckAssistBot, xAI-SearchBot, Grok-DeepSearch, Meta-ExternalFetcher, Kimi-User, Qwen-User, Amazon-Bedrock-AgentCore-Browser, Shap-User

​
Search index

OAI-SearchBot, Claude-SearchBot, PerplexityBot, Google-InspectionTool, Googlebot, MistralAI-Index, Bingbot, msnbot, Amzn-SearchBot, meta-webindexer, Kimi-SearchBot, TikTokSpider, Baiduspider, YouBot, Petalbot, YisouSpider, Sogou web spider, Coveobot, amazon-kendra, Algolia Crawler, ref-tools-crawler, ShapBot, AteveSearchSourceUrlDiscovery, YandexBot, YandexMobileBot, YandexAccessibilityBot, YandexRenderResourcesBot, YandexImages, YandexNews, YandexWebmaster, Yeti/, 360Spider, SeznamBot, Exabot

​
Training

GPTBot, ClaudeBot, GoogleOther, Google-Extended, Google-CloudVertexBot, Applebot-Extended, Applebot, Amazonbot, Meta-ExternalAgent, KimiBot, Bytespider, ERNIEBot, QwenBot, ChatGLM-Spider, DeepSeekBot, Deepseek, cohere-ai, cohere-training-data-crawler, AI2Bot, AI2Bot-Dolma, CCBot, PanguBot, ClueWeb-Crawler

​
Other AI

OAI-AdsBot, GrokBot, xAI-Bot, xAI-Grok, xAI-Web-Crawler, Grok, meta-externalads, facebookexternalhit, FacebookBot, Facebookbot, Doubaobot, YiyanBot, TongyiBot, AliyunBot, diffbot, FirecrawlAgent, SemrushBot-OCOB, SERankingBacklinksBot, Omgili, Omgilibot, webzio-extended, ImagesiftBot, VelenPublicWebCrawler, TurnitinBot, Timpibot, ICC-Crawler, DataForSeoBot, AwarioBot, AwarioSmartBot, AwarioRssBot, Kangaroo Bot, Sentibot, img2dataset, Meltwater, Seekr, peer39_crawler, Scrapy, Cotoyogi, aiHitBot, Factset_spyderbot, Barkrowler, HubSpot Crawler, AhrefsBot, SemrushBot, SiteAuditBot, AdsBot-Google, YaDirectFetcher, Google-Adwords, Screaming Frog, DotBot, MJ12bot, BLEXBot, serpstatbot, SEOkicks, trendictionbot, archive.org_bot, Decimal Bot, WebThreeCrawler, scopy-docs-crawler, quigon-manifest-fetcher, HeadlessChrome, PhantomJS

​
Excluded internal traffic

Requests whose user agent contains mintlify-revalidation or Mintlify-LinkChecker come from Mintlify’s own infrastructure and are removed before any counting.