Skip to content
Crawler identity

SergeBot

SergeBot is the crawler used by Serge’s website inspection tools. This page documents that technical service, including its user-agent and opt-out controls. Website inspection is separate from ChatGPT campaign preparation and does not guarantee ad eligibility or delivery.

User-Agent string

Every request from SergeBot identifies itself with this exact User-Agent:

SergeBot/1.0 (+https://www.serge.ai/bot; product-findability-scanner)

What SergeBot does

SergeBot only runs when a user initiates a scan. It is not an autonomous crawler and does not spider your site. Each scan makes a small, fixed set of requests to well-known paths:

ResourcePurpose
/llms.txtLLM product description
/llms-full.txtExtended LLM documentation
/openapi.jsonOpenAPI specification
/.well-known/agent.jsonA2A agent card
/.well-known/agents.jsonAgent directory card
/robots.txtCrawler permissions
/sitemap.xmlSite structure
/Homepage (structured data, JSON-LD)
/docs, /api, /developersDeveloper hub detection
/pricingPricing page detection

SergeBot also queries external registries (MCP Registry, PulseMCP, npm) for SDK and MCP server presence. These requests do not touch your infrastructure.

What SergeBot does not do

×
Scrape, index, or store your page content
×
Follow links or crawl beyond the fixed resource set
×
Train AI models on your content
×
Access authenticated or private pages
×
Bypass CAPTCHAs or bot protection challenges
×
Spoof its identity or rotate User-Agent strings

Rate limits

LimitValue
Max requests per domain per scan~20
Max concurrent requests per domain6
Per-request timeout8 seconds
Scan completes in< 30 seconds

SergeBot respects robots.txt directives and Crawl-delay values.

Controlling access

Allow SergeBot
User-agent: SergeBot
Allow: /
Block SergeBot
User-agent: SergeBot
Disallow: /

Blocking SergeBot can make its website checks inconclusive. Other crawlers have their own identities and rules; this setting does not determine OpenAI account approval or ad delivery.

Data handling

StoredScan scores, check results (pass/fail/warn), and fix-it suggestions
Not storedRaw page content, HTML source, or any data beyond what the checks need
RetentionScan results are kept for benchmark calculations. No PII is collected from scanned domains.

Contact

Questions about SergeBot, false positives, or access issues: