API Evangelist Scores 83% Agent-Ready — And the Last 17% Doesn't Apply

Key Takeaways

• API Evangelist scores 83, achieving Agent‑Native level on Cloudflare scanner
• Robots.txt, sitemap, and Link headers all passed 100% discoverability
• Markdown content negotiation returns text/markdown in 143 ms
• Explicit allow list for 13 AI bots signals consent for training
• OAuth discovery endpoints missing, lowering overall agent‑readiness score

Pulse Analysis

The rise of generative AI has turned the web into a data source for autonomous agents, prompting a push for machine‑readable metadata. Cloudflare’s new isitagentready.com tool quantifies how well a site complies with emerging standards such as RFC 8288 link headers, RFC 9727 API catalogs, and the Content‑Signals draft. By assigning a numeric score and tiered level, the scanner gives developers a clear benchmark for making their content discoverable and consumable by AI agents, a capability that was previously fragmented across disparate specifications.

API Evangelist’s implementation showcases a practical roadmap. The site earned perfect marks for discoverability by exposing a well‑structured robots.txt, sitemap, and Link headers that reference an API catalog and markdown alternatives. A Cloudflare Worker adds content negotiation, delivering markdown when an Accept: text/markdown request arrives, and the site openly lists 13 major AI bots in its robots.txt, accompanied by a Content‑Signal consent directive. These steps collectively earned an 83‑point score and a Level 5 Agent‑Native rating, signaling that the site is already usable by modern agents without additional crawling overhead.

However, the scan also revealed gaps that could hinder broader adoption. Both the OpenID configuration and OAuth‑authorization‑server endpoints returned 404, preventing agents from automatically discovering authentication flows. As AI agents increasingly require secure, authenticated access to APIs, missing OAuth metadata becomes a blocker. The findings underscore the need for unified, well‑documented endpoints and reinforce the industry’s move toward standardized agent‑ready signals. Providers that close these gaps will not only improve their AI‑readiness scores but also unlock new revenue streams from autonomous applications that rely on seamless, machine‑first interactions.