docs.cdp.coinbase.com
partial
Claude Code
intent

What does docs.cdp.coinbase.com do and who is it for? Explain it back to me.

5steps
38.8sduration
$0.0195cost
27,586tokens
6 steps1 reasoning step1 search
home
llms.txt
search
/developer-platform/pr…
/get-started/overview
60%
answer from your site
75%
answer efficiency
67%
followed site links
answer sources
  • from this site60%
  • from search results20%
  • from agent knowledge20%
url discovery
  • given in the task25%
  • followed a link50%
  • found via web search25%
insight

The agent assembled a partial but substantive overview of CDP by combining homepage content, a documentation index, and web search results, but could not verify pricing claims and found no competitor comparisons on the docs site itself. The site's public docs lack pricing and positioning information, forcing reliance on search snippets and external announcements.

  • ›The docs homepage and get-started pages are routing documents only—they do not explain CDP's purpose, pricing, or positioning; the agent had to construct the product summary from an llms.txt index and infer it from scattered product mentions across the docs.
  • ›Pricing information is entirely absent from docs.cdp.coinbase.com; the agent obtained all pricing facts from a web search that surfaced coinbase.com pricing pages and SDK type references, but could not fetch the main pricing page (403 Forbidden), so pricing claims are unverified.
  • ›The docs site publishes no comparison page or positioning statement versus alternatives; the agent acknowledged this gap explicitly rather than filling it from model memory, correctly limiting its claims.
  • ›The llms.txt index provided crucial product-area inventory (wallets, payments, x402, onramp, staking, etc.) that the homepage's broken product-area lists failed to render, making it the most useful discovery tool on the site.
  • ›Terminology is inconsistent across docs (Coinbase Wallet, CDP Wallets, Coinbase App, Server Wallets, Embedded Wallets, Agentic Wallet) and the docs do not clearly separate them by use case or custody model, creating a disambiguation burden.

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 →