Ora

Ora is an agent experience ranking platform built by era labs that scans websites, MCP servers, and MCP App URLs and scores them on how well AI agents can discover, authenticate with, and use them. The tool is actively live, has scanned over 10,851 sites according to the homepage, and publishes a public leaderboard across categories including Agent Tools, AI & ML, Developer Tools, CRM, and more.

What It Is

Ora measures what its creators call "Agent Experience" (AX)—the counterpart to UX, but for AI agents rather than humans. Where traditional SEO and UX optimization targets human visitors, AX optimization targets the autonomous agents that increasingly act on users' behalf. Ora's scoring system evaluates any product across five layers: discovery, identity, auth and access, agent integration, and user experience. Scores run from 0–100 and map to letter grades (A+ ≥ 95, A ≥ 86, B ≥ 70, C ≥ 48, D ≥ 28, F < 28).

How the Scoring Architecture Works

Each of the five layers maps to a concrete step an agent takes when trying to use a product on someone's behalf:

• Discovery — Can an agent find the product in the registries it checks?
• Identity — Does the product clearly communicate what it does and when to use it?
• Auth and access — Can an agent authenticate and act without a human in the loop? (Weighted highest, per the about page, because an agent that can find and understand a product but can't get in has achieved nothing.)
• Agent integration — Are APIs, SDKs, and protocols available for agents to actually act?
• User experience — When a human needs to step back in—for a payment, confirmation, or final decision—can the product surface real UI?

Ora runs static checks against docs, llms.txt, registries, and public APIs, then spawns real agents across platforms including ChatGPT, Claude, and OpenClaw that attempt to onboard and use the product end to end.

API and MCP Access

Ora exposes a full REST API and an MCP server, both free and rate-limited by IP at 10 requests per minute. Key endpoints include:

• POST /api/scan — runs a full agent-readiness scan and returns score, grade, and layer breakdown
• GET /api/score/{domain} — retrieves the most recent cached scan result
• GET /api/discover — finds the most agent-ready products for a described intent
• GET /api/feedback/{domain} — returns feedback submitted by AI agents about their experience using a product
• GET /api/badge/{domain} — returns an embeddable SVG badge showing the domain's score and grade

Agent feedback submission is exclusively available via MCP and is verified through HATCHA, described in the docs as "a reverse CAPTCHA that proves the caller is an agent."

Update: Deep Scan v1.1

The blog post dated May 13, 2026 describes Deep Scan v1.1 as a significant methodology update. According to the post, v1 ran real agents at every scoring layer; v1.1 makes those agents work harder. Discovery is now described as a true AEO/GEO benchmark across answer engines, and MCP scoring is graded against Anthropic's own best-practice guidelines rather than just checking whether the endpoint responds. The post notes that some scores will drop as a result of the stricter methodology.

Who It Is For

Ora is aimed at product and engineering teams at companies that want their products to be usable by AI agents—particularly those building or exposing MCP servers, REST APIs, or developer-facing tools. The about page frames the core audience as businesses that have spent years optimizing for human visitors and now need to understand a new set of requirements driven by agent-mediated access. The contact page explicitly invites outreach about scoring methodology, API access, and partnerships.