Agent Interface Platform

The complete stack for agent-native websites

AgentReady gives every website a structured interface for AI agents: a hosted MCP server, spec-compliant llms.txt, cited RAG answers, and analytics that show exactly what agents ask — and where your content fails them.

Index my site — free View pricing

All features

Everything your site needs to work with AI agents

Discovery

Hosted MCP server

A live Streamable HTTP MCP endpoint at agentready.it.com/api/mcp — no setup, no API key. Any MCP-compatible client (Claude, Cursor, Windsurf) connects immediately. 8 tools: ask_site, list_sites, get_site_capabilities, plan_site_action, submit_site, refresh_site, rate_answer, get_site_facts.

https://www.agentready.it.com/api/mcpDocs →

Discovery

llms.txt generation

AgentReady generates a spec-compliant llms.txt for every indexed site, served at your site's canonical URL. AI agents discover and respect it automatically — no DNS change, no code, no plugin. Full-content llms-full.txt also generated.

/api/sites/{siteId}/llms.txt

Retrieval

Cited RAG answers

When agents query your site, AgentReady retrieves exact quotes and source URLs from your real content. Answers include matched page URLs, freshness metadata, and similarity scores — no hallucinations, no invented version numbers.

POST /api/sites/{siteId}/ask

Observability

Agent analytics

See every agent query, which AI clients sent it, success rate, and p50/p95 latency. The analytics dashboard breaks down top agents, top questions, and answer quality over 7, 30, or 90 days depending on plan.

/analytics/{domain}Docs →

Quality

Agent Simulator

Run 5 representative agent tasks against your site and get a scored readiness report before you launch. The simulator runs in a real agent environment and catches gaps that manual testing misses.

Pro feature

Quality

Query gap detection

AgentReady tracks every question that returned no confident answer. Pro users see the full gap list and can generate an AI-drafted content brief for each — structured prompts for site owners to fill in with real facts, never fabricated answers.

Pro feature

Trust

Content trust scanning

Every crawled page is hash-pinned (SHA-256) and scanned for prompt injection: CSS-hidden AI-directed instructions, comment-embedded fake system prompts, and visible manipulation text. Flagged content is stripped before indexing, never reaching the RAG prompt.

All plans

CI/CD

GitHub check runs

On every site refresh, AgentReady re-runs your tracked queries and posts the answer quality score as a GitHub Check on your HEAD commit. Regressions fail the check automatically — so a docs deploy never silently breaks agent answers.

Pro feature

Quality

Answer tests (expected-fact CI)

Define test queries with expected facts — e.g. "Pro plan costs $49/month." After every refresh, each fact is evaluated against the live answer. Tests pass, partial, or fail with an explanation. Think of it as snapshot testing for your AI-served documentation.

POST /api/sites/{siteId}/tests/run

CI/CD

Content Push API

Push individual pages to the index immediately after deploy — no need to re-crawl the whole site. Accepts URL + content via REST or Bearer API key. Add it to your docs deploy pipeline: changed pages are indexed in seconds.

POST /api/sites/{siteId}/push

Trust

Structured fact attestation

AgentReady extracts key-value facts from your site on every crawl — pricing, limits, features, contact info. Pro owners attest facts as owner-verified ground truth. Attested facts are injected before scraped content in every AI answer, so agents always get the correct answer directly from the source.

GET /api/sites/{siteId}/facts

Observability

Query intent analytics

Every agent query is classified into one of six intent buckets — pricing, comparison, technical, support, evaluation, or general — in real time, at zero cost. The analytics dashboard breaks down which intent types dominate your traffic so you know exactly what agents are trying to decide.

/analytics/{domain}

Observability

Site changelog

Every crawl generates a diff: new pages added, existing pages modified, removed pages flagged. The changelog lets you verify that a deploy actually changed what you expected — and shows agents when your content last changed, helping them calibrate answer freshness.

GET /api/sites/{siteId}/changes

Trust

Agent registry

Agents self-register with a name, description, contact email, and user-agent pattern. Registered agents get a verified badge in your analytics — so you can tell the difference between a Claude-powered agent built by your customer and an unknown scraper hitting your endpoint.

GET /api/agents

Quality

Site freshness alerts

Pro sites that haven't been re-indexed in 7 days get a staleness email with a one-click refresh link and the exact CLI command. Agents querying stale docs get wrong answers — freshness alerts close that loop before your customers notice.

Pro feature

For site owners

Your site, agent-ready in 60 seconds

No code, no developer, no API key. Paste any public URL and AgentReady handles the rest.

Paste your URL

Sign in free and paste any public URL. AgentReady crawls and indexes your content in under 60 seconds — pages, docs, product descriptions, everything.

Agents connect automatically

Your site gets an MCP endpoint, llms.txt, capability manifest, and cited-answer API. Any agent can query it immediately from any MCP client.

Monitor and improve

See every agent question in your analytics dashboard. Spot content gaps, catch answer regressions with GitHub checks, and get AI-drafted briefs for missing content.

For agent developers

One MCP endpoint for any site on the web

Connect once, query any indexed site. No per-site setup, no separate API keys. Compatible with Claude Desktop, Cursor, Windsurf, and every MCP client.

Remote MCP endpoint (Streamable HTTP)

https://www.agentready.it.com/api/mcp

stdio bridge (local clients)

npx @agentreadyweb/mcp

CLI commands

npx @agentreadyweb/mcp grade yoursite.comnpx @agentreadyweb/mcp ask yoursite.com "How do I authenticate?"npx @agentreadyweb/mcp refresh yoursite.com

Use refresh in CI after deploys to keep agent answers current.

Full client setup guide →

8 MCP tools

ask_siteCited answers from any indexed site — auto-indexes on first query
list_sitesBrowse all indexed sites with their endpoints and summaries
get_site_capabilitiesInspect a site's capability manifest and agent-readiness grade
plan_site_actionCreate a grounded, bounded read-only plan from site content
submit_siteExplicitly index a new public site (~60 seconds)
refresh_siteForce a re-crawl and re-index of any site
rate_answerSubmit quality feedback (rating 1–5) to improve future answers
get_site_factsReturn structured facts (pricing, limits, features, contact) with owner-verified ground truth flagged

Browse the indexed site directory →

Integrations

Fits into the tools you already use

GitHub check runs

Pro

Every refresh posts a quality score to your HEAD commit. Regressions block the merge.

Slack regression alerts

Pro

Get a Slack message the moment agent answer quality drops below your threshold.

Private MCP access

Pro

Lock your MCP endpoint behind an API key — for internal knowledge bases or pre-release content.

Webhook refresh

All plans

Hit /api/sites/{siteId}/refresh from your CI pipeline to re-index after every deploy.

Agent-ready badge

All plans

Embed a live badge on your site showing agent-readiness grade. Updates after every crawl.

Ownership verification

All plans

Verify domain ownership with a meta tag. Unlock management features without sharing credentials.

Pricing

Start free, scale as you grow

Free to start — no credit card required. Upgrade for more sites, longer analytics, and CI-grade quality assurance.

Free

$0forever

✓1 site · 100 pages
✓5 refreshes/month
✓Public MCP + llms.txt
✓Cited RAG answers
✓7-day analytics

Pro

$49/month

✓5 sites · 500 pages each
✓50 refreshes/month
✓Private MCP + API keys
✓GitHub checks + Slack alerts
✓30-day analytics
✓Agent Simulator
✓Query gap detection
✓Answer tests (expected-fact CI)
✓Content Push API

Team

$149/month

✓Unlimited sites + pages
✓Unlimited refreshes
✓All Pro features
✓90-day analytics
✓Early access

See full plan comparison, FAQs, and Enterprise →

Ready to make your site
work with every AI agent?

Free to start. Paste a URL and get a hosted MCP server, llms.txt, and cited answers in under a minute.

Index my site — free Talk to us about Enterprise

The complete stack for agent-native websites

Everything your site needs to work with AI agents

Hosted MCP server

llms.txt generation

Cited RAG answers

Agent analytics

Agent Simulator

Query gap detection

Content trust scanning

GitHub check runs

Answer tests (expected-fact CI)

Content Push API

Structured fact attestation

Query intent analytics

Site changelog

Agent registry

Site freshness alerts

Your site, agent-ready in 60 seconds

Paste your URL

Agents connect automatically

Monitor and improve

One MCP endpoint for any site on the web

Fits into the tools you already use

GitHub check runs

Slack regression alerts

Private MCP access

Webhook refresh

Agent-ready badge

Ownership verification

Start free, scale as you grow

Free

Pro

Team

Ready to make your sitework with every AI agent?

Ready to make your site
work with every AI agent?