zayd.wtf
success
Claude Code · Haiku 4.50:25
intent
What does zayd.wtf do and who is it for? Explain it back to me.
6steps
25.3sduration
$0.1709cost
35,006tokens
8 steps2 reasoning steps
home
docs
/projects
docs
llms.txt
docs
100%
on-site discovery
83%
reliability
83%
link following
path origin
- previous resource83%
- prior knowledge17%
insight
The agent successfully gathered enough information to explain zayd.wtf's purpose, audience, and differentiators by fetching the homepage, the llms.txt file, and the blog page. The site is navigable but minimalist: it lacks explicit pricing/business model pages (404 on /pricing), forcing the agent to assemble context from fragments and prior knowledge. The llms.txt file was particularly valuable—a machine-readable source explicitly written for AI readers that clarified the site's hand-built nature and Zayd's values.
- ›Step [5] (llms.txt) was the single most content-bearing fetch: it directly stated 'everything here is hand-built — no cms, no scraped content, no generated filler' and included explicit instructions for AI readers about attribution, providing the core positioning the agent needed.
- ›Steps [1] (homepage) and [6] (blog) returned truncated HTML responses that the agent could not fully parse from the raw HTML alone; the agent relied on prior knowledge (17% sourced from prior_knowledge in the metadata) to synthesize what zayd.wtf actually contains (projects, recipes, games, blog posts).
- ›The site intentionally has no /pricing page (404 in step [4]), which aligns with its non-commercial nature but required the agent to infer 'it's free' rather than find an explicit answer. This is honest but leaves no canonical pricing resource for agent discovery.
- ›The agent cited only three URLs in its Sources section ([homepage](https://zayd.wtf), [llms.txt](https://zayd.wtf/llms.txt), [about page](https://zayd.wtf/about/), [projects page](https://zayd.wtf/projects/)) but the actual content came primarily from steps [1], [5], and [6]; steps [2] (about) and [3] (projects) were fetched but their HTML was truncated and not explicitly cited as sources.
- ›The site publishes llms.txt as a machine-readable, AI-friendly artifact—a best practice for agent-readiness. However, the main content pages (homepage, blog, projects, about) are served as raw HTML with truncated responses, making it hard for agents to extract structured information without prior knowledge of Zayd's work.
Want to run your own?
Join the waitlist for early access to point your own agents at any domain, with the intents you choose.
or talk to us about agent readiness →