docs.knock.app
success
Claude Code
intent

What does docs.knock.app do and who is it for? Explain it back to me.

15steps
79.2sduration
$0.8644cost
211,214tokens
18 steps3 reasoning steps5 searches
home
/overview
home
/guide
/getting-started
docs
search
/getting-started/what-…
/getting-started/knock…
search
llms-full.txt
search
docs
search
search
70%
on-site discovery
70%
reliability
50%
link following
path origin
  • previous resource50%
  • web search30%
  • prior knowledge20%
insight

The agent successfully assembled a comprehensive explanation of Knock's purpose, audience, pricing, and competitive positioning by combining documentation fetches with web search results and prior knowledge. The site's documentation was navigable once discovered via search, but the homepage and direct URL guessing failed, requiring the agent to use web search as the primary discovery mechanism. The agent retrieved accurate core information but had to work around gaps in pricing detail and feature comparison matrices that weren't explicitly published in machine-readable form.

  • ›Steps [9] and [10] returned the core identity content: 'What is Knock?' and 'Knock's features' pages from docs.knock.app/getting-started/, which explained the three main products (Workflows, Broadcasts, Guides) and the cross-functional audience. Step [12] (llms-full.txt) contained a similar overview in text format, demonstrating the docs are structured for machine consumption.
  • ›The agent could not discover these pages through site navigation—direct URL guesses at /overview, /getting-started, and /guide all returned 404s (steps [2], [5], [6]). Discovery required web search (step [8]) to find the correct /getting-started/what-is-knock and /getting-started/knocks-features URLs, indicating the homepage and sitemap do not surface these entry points via direct traversal.
  • ›Pricing details (step [14], knock.app/pricing) were vague in machine-readable form; the agent had to infer pricing tiers from search results and G2 reviews rather than from a structured pricing table. The agent explicitly flagged missing information: specific feature matrices by plan, API rate limits, migration timelines, and whether channels have per-unit costs—gaps that reflect incomplete documentation or landing page design rather than site navigation failure.
  • ›The site's foundation is agent-ready (docs.knock.app serves structured content, llms-full.txt exists, SEO metadata is present), but discoverability is weak: a direct visitor to docs.knock.app/ would struggle to find 'What is Knock?' without search or prior knowledge. The friction outcome 'succeeded_with_light_bridge' reflects this—web search had to bridge the gap between the homepage and the actual content pages.

See every agent that visits docs.knock.app

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 →