kernal.sh
success
Claude Code
intent

What does kernal.sh do and who is it for? Explain it back to me.

11steps
51.5sduration
$0.5907cost
113,666tokens
19 steps8 reasoning steps4 searches
home
docs
/index.md
docs
docs
docs
search
docs
search
search
search
100%
on-site discovery
100%
reliability
86%
link following
path origin
  • previous resource86%
  • prior knowledge14%
insight

The agent gathered a comprehensive understanding of Kernel's core offering (browser infrastructure for AI agents), pricing model (usage-based, ~$0.10–$0.30/min), target audience, and key differentiators (speed, stealth/bot-evasion, browser pools) by combining markdown documentation, web search results comparing alternatives, and prior knowledge. However, the site's pricing details were scattered and incomplete—exact tier limits, free-tier allowances, and feature breakdowns were not clearly documented in a single location, forcing the agent to infer and assemble information from fragments.

  • ›Step [6] returned the core value proposition via markdown (index.md): 'KERNEL gives web agents cloud-based browser access for fast, scalable automation' with mentions of isolated sessions, profiles, proxies, and observability—this was the primary content source for what-they-do and who-it's-for.
  • ›Step [7] (llms.txt) provided the documentation structure and navigation hints but was largely an index; the agent used web search (steps [14] and [15]) to find comparison content and feature details (stealth mode, proxy types, bot detection) not prominently surfaced on the main domain.
  • ›Step [12] (docs/info/pricing page) returned HTML but the response was truncated and agent-unreadable; the agent could not extract concrete pricing tiers or limits from the actual page fetch, relying instead on web search results and prior knowledge to estimate '$0.10–$0.30 per minute.'
  • ›Pricing information was the weakest point: the agent explicitly noted 'I couldn't find the exact tiers and limits documented in an obvious place' and 'free tier limits unclear.' The site publishes a pricing page but does not expose it in machine-readable or easily parseable form.
  • ›The site's agent-readiness is mixed: documentation is extensive and technical (good for identity), but marketing copy and pricing are rendered as client-side JavaScript (poor for direct fetch), and structured comparisons against competitors are absent, forcing the agent to synthesize information from external sources.

Help the next agent finish on kernal.sh

Front builds the agent version of your site and shows which agents visit, what they came for, and what your pages couldn't answer.

or talk to us about agent readiness →