# AI Agent Readiness Scanner

> Paste a website, REST API or MCP endpoint and get a 0 to 100 score for AI agent readiness, a letter grade, and a fix prompt for every failed check.

Canonical URL: https://www.startuphub.ai/agent-readiness
Free to use: yes. Publisher: StartupHub.ai

## What it offers

- It runs about 40 checks across six categories: discoverability, content, access control, capabilities, commerce and quality.
- Each check returns pass, fail, warn or skip, and a warning counts as 40 percent of that check weight while a skip does not count.
- It probes /robots.txt, /sitemap.xml, /llms.txt, /openapi.json, /.well-known/mcp/server-card.json and /.well-known/ap2.
- The commerce category checks x402 (at /x402 and /.well-known/x402), MPP, AP2 and ACP (at /.well-known/acp) separately, and a site without paid endpoints is not penalised.
- Every failed check comes with a paste-ready fix prompt for a coding agent such as Claude or Cursor.

## Why it is different

- Live Agent Run: real agent frameworks Claude Code, Codex, OpenCode, Cursor, Gemini CLI and a plain model loop as control answer seven questions, eight when commerce applies.
- The full scan fetches the site as 17 named AI clients and crawlers, including Claude-User, ChatGPT-User, Perplexity-User, GPTBot, ClaudeBot, Googlebot and Bingbot.
- llms.txt is checked against the v2 shape: an H1, a summary and described links.
- Grades: A+ is 90 or above, A 80, B 70, C 60, D 40, F under 40.

## How to use it

1. Paste a URL: your own site, a competitor, a REST API or an MCP endpoint.
2. Optionally switch on the Live Agent Run.
3. Run the scan and read the overall score, the grade and the six category scores.
4. Open each failed check and copy its fix prompt into Claude or Cursor.
5. Apply the fixes and scan again to confirm the new score.

## Inputs and outputs

- Inputs: A public URL or domain, with an optional ?url= parameter for a shareable report link.
- Outputs: A 0 to 100 score, a letter grade, six category scores, every check with its fix prompt and Live Agent Run results per framework, as a page or as JSON via the API.

## Limits and plans

- Without an account, the basic scan is free with no signup and allows up to 10 scans a day from one network.
- On the Free plan, a full scan with the Live Agent Run needs a free account and includes 3 a month, while every paid plan is unlimited with no card required.
- The REST endpoint costs 1 credit per call and the Free plan includes 1,500 API credits a month in one pool for REST and MCP.
- With x402, agents can pay per call at $0.001 USDC with no API key.
- Plans and prices: https://www.startuphub.ai/pricing

## API and MCP

| Method | Endpoint | Auth and cost |
|---|---|---|
| POST | `https://www.startuphub.ai/api/v1/agent-readiness` | Bearer key, 1 credit; or x402 $0.001 with no key |

MCP server: `https://www.startuphub.ai/api/v1/mcp`
- `scan_agent_readiness` (no key)

API keys and docs: https://www.startuphub.ai/api-docs

## Related tools

- [API and MCP Security Scanner](https://www.startuphub.ai/api-security): Point it at an API or MCP endpoint and see what it gives a stranger with no credentials, graded A+ to F with a fix for every finding.
- [Tech Stack Checker](https://www.startuphub.ai/tech-stack): Paste a URL to see the hosting, frameworks, analytics, payments, auth and DNS behind any site linked to the company that runs it.
- [Website Speed Test](https://www.startuphub.ai/website-speed-test): A full Lighthouse run through Google PageSpeed Insights for any URL, with lab and real-user field data side by side, on mobile or desktop.
- [All free tools](https://www.startuphub.ai/free-tools)