Ora's Ranker 3.0 grades whether agents can actually use your site; scores drop 40+ points
yenkel · x · 2026-11-12
Ora shipped Ranker 3.0, rescoring sites on whether AI agents can actually complete tasks. With personal agents like Muse and Instinct now booking and buying for people, merely having an MCP server or WebMCP tools no longer earns points—each tool is graded through checkout.
- Among 22,492 sites scoring 60+ on 2.0, the typical score fell from 67 to 48.
- Sites with a working MCP server and documented APIs lose 6 points; sites with neither lose 26.
- Some sites dropped over 40 points.
- The rubric tracks protocol evolution: WebMCP irreversibility hints, multi-tool MCP servers, MCP Apps in chat. "Outcomes > tech"—what matters is how hard it is for an agent to do the job.
More from coding & agent
- Stanford NLP course's best homework: build your own agent harness — stanfordnlp · 2026-11-12
- Anthropic details agent memory: dreaming organizes Claude's memory, wiki+search hits 61% — RLanceMartin · 2026-11-12
- Talk preview: scaling 1,000 Fable 5.2 agents to formalize Fermat's Last Theorem — RLanceMartin · 2026-11-12
- Opinion: AI infra is becoming a coding runtime as intelligence units outgrow tokens — prateekj · 2026-11-12
- Claude Docs/Slides block external MCP clients: no public API, export only — steipete · 2026-11-12
- Anthropic explains agent memory: 'dreaming' lets Claude consolidate session transcripts into notes — rseroter · 2026-11-12