Hosted proof,
local control.
Hosted journeys on Pro run 10 managed personas with browser verification. Local BYOK journeys run from the CLI with your own model keys, so developers can verify the same shopping paths without a hosted run.
One run. Two layers of evidence.
Hosted agent journeys
Ten managed personas make real storefront requests across discovery, product evaluation, returns, edge cases, security, mobile, and checkout handoff.
- CoverageDiscovery, product detail, cart integrity, returns, security, mobile, and checkout readiness
- ModelsGPT, Claude, and Gemini — cross-model evidence from the production stack
- OutputPer-persona scores, journey reports, blockers, and remediation recommendations
Browser evidence
When your store exposes browser agent tools, the same run adds a live Chrome replay with screenshots and tool traces so you can verify the storefront works the way it should.
- EngineCloudflare Browser Rendering with live storefront navigation
- RequiresBrowser agent tools on your store (Fix can generate them)
- OutputJourney screenshots, tool invocation logs, and setup guidance when tools are missing
Five stages of an agent shopping flow
Each stage maps to a real step in an AI agent's purchase flow.
JOURNEY FUNNEL
18% conversion18% of agents complete a purchase. The biggest drop happens at Browse → Evaluate — most agents can't parse product schema.
Works from CLI, API, and MCP
AI agents discover and invoke tools programmatically. Developers get the same access from the command line.
# Hosted agent journeys (live storefront evidence)curl -X POST https://agenticcom.ai/api/v1/test/run \ -H "Authorization: Bearer col_..." \ -d '{"url": "https://my-store.com"}' # Results include: personas[], summary,# journey_reports, remediations[],# per-model scores, total_cost_usdInstall: npx -y @getcolter/cli · MCP: npx -y @getcolter/cli mcp --admin-tools
Common questions
What happens during a hosted journey?
10 managed personas attempt shopping journeys on your store across Claude, GPT, and Gemini workflows. Each one checks a different path: discovery, product detail, cart integrity, checkout readiness, returns, security, mobile, and edge cases.
Will a journey affect my store?
No purchases are completed. Personas browse, add to cart, and attempt checkout but stop before payment. Read-only except for cart actions.
How is this different from Check?
Check scans public evidence. Hosted journeys show what happens when managed agents try to use the store. Check tells you what is missing; journeys show where runtime paths break or need better tool support.
What do I get in the report?
Per-persona scores, journey narratives, remediation recommendations, and a PDF export. You see where each managed journey succeeded, failed, or needed runtime support the storefront has not exposed yet.
Do hosted journeys include browser verification?
Yes. When your store exposes browser agent tools, the same run adds a live browser replay with screenshots and tool traces for verification.
Monitor what happens in production.
Hosted journeys verify shopping paths before launch. Lens watches real agent traffic - who visits, what they try, where they drop off.
Learn about Lens