Skip to main content
Check your store

What can agents see?

Paste your store URL. Colter probes 30+ signals and shows what AI agents can find, what they can't, and what needs a live test.

Free · No account required · 30-second scan

See a sample report

Read-only. No admin access. Results in seconds.

Colter Check
Free

Instant readiness scan.
Free, always.

Paste a URL. Colter probes public protocol manifests, structured data, buying surfaces, and WebMCP tools — then separates verified evidence from gaps that need runtime testing.

Scan your store
colter check
$colter check https://my-store.com
Scanning protocols, structured data, agent surfaces...
✓ JSON-LD Product detected
✓ UCP manifest valid
✓ WebMCP tools (3 found, checkout tools)
✗ Missing llms.txt
✗ robots.txt blocks GPTBot
Score: 72/100 (B) • 5 fixes available
What gets checked

30+ signals across 5 dimensions

Every scan covers the public surfaces AI agents use to discover, evaluate, and prepare to buy from a store.

  • DiscoveryJSON-LD, sitemap, llms.txt, robots.txt, AGENTS.md, WebMCP tools
  • TransactionUCP, ACP, MCP endpoints, checkout flow accessibility, payment surfaces
  • SecurityHSTS, CSP, X-Frame-Options, rate limiting, agent auth support
  • EcosystemProtocol coverage breadth — UCP, ACP, MCP, WebMCP, A2A agent card
  • ContentJSON-LD field completeness, OG tags, SSR readiness, product data quality
  • WebMCPTool count, checkout tools, description quality, parameter coverage (0-26 score)
Who it's for

Works for merchants, agencies, and developers

Merchants

  • Scan in 30 seconds, no technical knowledge needed
  • Plain-English verdict: what works, what doesn't, what to fix first
  • Share results with your dev team or agency

Agencies

  • Scan client stores for upsell opportunities
  • Batch scan a portfolio to prioritize work
  • Quantify the gap between current state and agent-ready

Developers

  • CLI and API for CI/CD integration
  • MCP tool for AI-assisted development workflows
  • Detailed protocol-level diagnostics
For agents and developers

Works from CLI, API, and MCP

AI agents discover and invoke tools programmatically. Developers get the same access from the command line.

CLI
colter check
$colter check https://allbirds.com
$colter check https://allbirds.com --json | jq .composite_score
72
$colter check https://allbirds.com --format plain
REST API
REST APIbash
curl -X POST https://agenticcom.ai/api/v1/check \
  -H "Content-Type: application/json" \
  -d '{"url": "https://my-store.com"}'
 
# Returns: verdict, composite_score, grade,
# protocols, web_signals, recommendation,
# plain_language, webmcp_quality_score

Install: npx -y @getcolter/cli · MCP: npx -y @getcolter/cli mcp --admin-tools

FAQ

Common questions

What does Check actually scan?

Read-only scan of HTML, headers, robots.txt, structured data, and protocol endpoints. No admin access needed, safe for production.

How long does a scan take?

30 seconds on average. Results include a score, dimension breakdown, and prioritized fix plan.

Is Check free?

Yes, always. 3 scans per day on the free tier. No account required for your first scan.

What protocols does Check look for?

UCP (Google), ACP (ChatGPT), MCP (Claude), WebMCP (browser agents), plus JSON-LD, llms.txt, and AGENTS.md.

What's a good score?

75+ means strong public readiness evidence. Runtime cart, checkout, and payment behavior should still be validated with agent journeys before stronger claims.

Next step

Found gaps? Fix them.

Check shows what's missing. Fix generates a platform-specific preview, ranked by estimated readiness lift.

Learn about Fix