Vercel Introduces ‘Is Agentic’, a Free Agent-Readiness Scoring Tool That Audits Public Websites Using Ora’s 100+ Checks

By

Vercel has released Is Agentic, a public tool that scores how readily AI agents can discover, access, understand, and use a website. Scans are run and scored by Ora, an agent-experience research company from era labs. Vercel operates the interface, report pages, storage, and the grouping that produces the displayed score.

Is it deployable?

Yes — at zero cost. Is Agentic is currently free, with no paid plans, subscription charges, or per-report fees. The public site, read-only API, CLI, and MCP server require no API key or billing account. Enter a URL in the browser, or run the CLI:

Copy Code
npx is-agentic <domain>
npx is-agentic <domain> --json

Which level of company: Any organization owning a public web surface. Seed-stage startups get a free baseline audit with no procurement. Mid-market SaaS teams can wire --json output into CI as a regression gate. Enterprises can benchmark documentation portals, developer sites, and commerce surfaces across business units.

Industries: Developer tools and SaaS, e-commerce and retail, travel and hospitality, fintech, marketplaces, healthcare provider directories, and media or documentation publishers. Anywhere an agent may shop, book, integrate, or cite on a user’s behalf.

Applications: Pre-launch agent-readiness audits. CI checks that fail builds when server-rendered content regresses. Competitive benchmarking against public scores. Validating an MCP server or OpenAPI surface is actually discoverable. Prioritizing technical fixes with copy-pasteable remediation prompts.

What Exactly it Measures

Ora’s methodology walks four layers an agent moves through. Discovery carries 20 points across 15 checks. Access carries 30 points across 41 checks. Usability carries 40 points across 56 checks. Payments carries 10 points across 6 checks. That totals 118 checks, matching Vercel’s “100+ checks” claim. Ora says the checklist is reverse-engineered from real agent runs, not authored by opinion.

Ora also publishes a letter scale: A+ at 95–100, A at 86–94, B at 70–85, C at 48–69, D at 28–47, and F at 0–27.

Vercel regroups those checks into its own displayed scoring model. Essential checks share an 80-point pool. Recommended checks share a 20-point pool. Emerging signals add a bonus capped at five points, and their absence never lowers a score. Not-applicable checks are excluded rather than counted as failures. Partial results receive proportional credit, and duplicated check IDs across MCP surfaces are averaged.

That applicability logic matters. Recommended checks activate only when scan evidence positively identifies an API, OAuth flow, GraphQL endpoint, MCP server, developer portal, or commerce surface. A marketing site is not penalized for omitting an interface it never claimed to offer.

Inside a report

Every finding ships with observed evidence and, where available, a concrete recommendation. A live API response shows the shape: score, score_label, scanned_at, eligible_checks, a score_breakdown with earned and available points per tier, and an issues array. Each issue carries id, name, details, recommendation, result, and tier.

Real check IDs include content-no-js, agent-friendly-404, markdown-negotiation-vary, json-ld, sitemap, trust-anchors, and metadata-completeness. The 404 check asks for a real HTTP 404 status rather than a 200 serving an app shell.

Reports also include an observed agent journey showing where one agent hit friction. Vercel keeps that run outside the numeric score, since one task cannot represent every agent. “Prompt to fix” converts actionable recommendations into an implementation brief for a coding agent.

Machine interfaces

Three read-only surfaces expose completed reports. The report API lives at /api/v1/report, rate-limited to 120 requests per client IP per 60-second window. Errors use RFC 9457 problem details with stable codes such as invalid_url, report_not_found, and rate_limit_exceeded. The OpenAPI description is the supported integration contract, discoverable through an RFC 9727 API catalog. Deprecations signal via RFC 9745 headers, with Sunset announced at least 90 days ahead.

The MCP server at https://is-agentic.com/mcp exposes is_agentic_get_report, is_agentic_get_methodology, and is_agentic_get_developer_docs over Streamable HTTP. Hosts supporting MCP Apps render an interactive score card. An official agent skill installs via npx skills add vercel-labs/is-agentic.

The site practices what it scores. Report pages render the score in the initial HTML response, serve a Markdown variant under Accept: text/markdown, and set Vary: Accept so shared caches never mix representations.

Key Takeaways

  • Is Agentic scores public sites on agent discovery, access, usability, and payments — free, with no API key.
  • Ora supplies 118 checks across four weighted layers; Vercel regroups them into Essential 80, Recommended 20, Bonus 5.
  • Not-applicable checks are excluded, so marketing sites are not penalized for missing APIs or commerce surfaces.
  • Reports ship as HTML, Markdown, JSON, CLI output, and MCP tools, all read-only and unauthenticated.
  • Vercel states it is not a certification, security audit, or accessibility review — treat it as a prioritized technical review.

The post Vercel Introduces ‘Is Agentic’, a Free Agent-Readiness Scoring Tool That Audits Public Websites Using Ora’s 100+ Checks appeared first on MarkTechPost.



from MarkTechPost https://ift.tt/rYCEAmg
via IFTTT

Comments

Post a Comment

Popular posts from this blog

Microsoft AI Proposes BitNet Distillation (BitDistill): A Lightweight Pipeline that Delivers up to 10x Memory Savings and about 2.65x CPU Speedup

By
Image
Microsoft Research proposes BitNet Distillation , a pipeline that converts existing full precision LLMs into 1.58 bit BitNet students for specific tasks, while keeping accuracy close to the FP16 teacher and improving CPU efficiency. The method combines SubLN based architectural refinement , continued pre training , and dual signal distillation from logits and multi head attention relations. Reported results show up to 10× memory savings and about 2.65× faster CPU inference , with task metrics comparable to FP16 across multiple sizes. What BitNet Distillation changes? The community already showed that BitNet b1.58 can match full precision quality when trained from scratch, but converting a pretrained FP16 model directly to 1.58 bit often loses accuracy, and the gap grows as model size increases. BitNet Distillation targets this conversion problem for practical downstream deployment. It is designed to preserve accuracy while delivering CPU friendly ternary weights with INT8 activa...
Read more

Nanbeige4-3B-Thinking: How a 23T Token Pipeline Pushes 3B Models Past 30B Class Reasoning

By
Image
Can a 3B model deliver 30B class reasoning by fixing the training recipe instead of scaling parameters? Nanbeige LLM Lab at Boss Zhipin has released Nanbeige4-3B, a 3B parameter small language model family trained with an unusually heavy emphasis on data quality, curriculum scheduling, distillation, and reinforcement learning. The research team ships 2 primary checkpoints, Nanbeige4-3B-Base and Nanbeige4-3B-Thinking, and evaluates the reasoning tuned model against Qwen3 checkpoints from 4B up to 32B parameters. https://ift.tt/bMCdZs2 Benchmark results On AIME 2024, Nanbeige4-3B-2511 reports 90.4, while Qwen3-32B-2504 reports 81.4. On GPQA-Diamond, Nanbeige4-3B-2511 reports 82.2, while Qwen3-14B-2504 reports 64.0 and Qwen3-32B-2504 reports 68.7. These are the 2 benchmarks where the research’s “3B beats 10× larger” framing is directly supported. The research team also showcase strong tool use gains on BFCL-V4, Nanbeige4-3B reports 53.8 versus 47.9 for Qwen3-32B and 48.6 for Qwen...
Read more

Technical Deep Dive: Automating LLM Agent Mastery for Any MCP Server with MCP- RL and ART

By
Image
Table of contents Introduction What Is MCP- RL? ART: The Agent Reinforcement Trainer Code Walkthrough: Specializing LLMs with MCP- RL Explanation: Under the Hood: How MCP- RL Generalizes Real-World Impact and Benchmarks Architectural Overview Practical Integration Summary Introduction Empowering large language models (LLMs) to fluidly interact with dynamic, real-world environments is a new frontier for AI engineering. The Model Context Protocol (MCP) specification offers a standardized gateway through which LLMs can interface with arbitrary external systems—APIs, file systems, databases, applications, or tools—without needing custom glue code or brittle prompt hacks each time. Still, leveraging such toolsets programmatically, with robust reasoning across multi-step tasks, remains a formidable challenge. This is where the recent combination of MCP- RL (a reinforcement learning loop targeting MCP servers) and the open-source ART (Agent Reinforcement Trainer) library ...
Read more