AI-ready commerce

Readable for assistants. Controlled for checkout.

OneNet Servers gives AI agents clean discovery surfaces for products, pricing, and domains while keeping payment, account, and service changes behind verified customer handoff.

Read-only product catalogLive domain searchVerified checkout handoff

Backed, Featured and Trusted by Top media, organisations and industry leaders

Amazon Web Services
Itana
NiRA
Tech Cabal
Tech Nation
British Business
Climate Neutral Data Center
Raenest Business
UK Home Office
Surface map

Three clean ways agents can work with OneNet Servers.

Pick the surface that matches the job: public pages for context, JSON APIs for structure, and MCP for tool-driven workflows.

Public pages

Agents can read pricing, plan comparisons, support guidance, and commercial pages exactly as a customer would.

Best for browsing, summarising, and citing public content.

Catalog API

Structured plan, pricing, and domain-search endpoints keep product discovery predictable for assistants and workflow tools.

Best for machine-readable comparisons and cart preparation.

MCP server

A JSON-RPC tool surface exposes product listing, domain search, and checkout requirements for agent runtimes.

Best for tool-based agent orchestration with consistent contracts.
How it works

The flow is designed to help agents without letting them overreach.

Discovery is fast, structured, and reusable. Checkout remains a verified customer handoff instead of a blind tool call.

01

Discover the right fit

Agents read plans, compare billing options, and narrow domain choices without changing customer state.

02

Prepare the decision

The app validates cart shape, domain requirements, and payment readiness before any customer details are finalised.

03

Hand off securely

Checkout, payment, account creation, and service changes stay behind verified customer handoff and secure checkout controls.

Production controls

Useful agent access without hidden billing risk.

These controls are what keep the public surfaces production-safe for AI agents instead of turning them into a fragile checkout shortcut.

No direct write path

Agents can discover and prepare. They cannot silently create orders, pay invoices, register domains, or change services.

Approval before action

Every state-changing step is gated behind customer context, explicit approval, and app-level validation.

Reconciliation after outcomes

Paid, unpaid, free-plan, trial, and included-domain outcomes are reconciled so the customer sees the right result afterwards.

Built for production agents

Read-only discovery is fast, structured, and predictable so agent workflows can stay useful without becoming risky.

Interface comparison

Choose the right surface for the task.

Pages, APIs, and MCP all expose the same discovery posture, but each fits a different level of automation.

Agent-safe interface map
MetricsPublic pagesHuman-readableCatalog APIStructured JSONMCP serverTool-firstFeatured
Live plan and pricing discoveryIncludedIncluded
Domain search without cart mutationManual flowIncluded
Structured machine outputLimitedIncluded
Direct write accessNot includedNot included
Best fitCustomer-facing browsingApps and automation

All state-changing actions stay verified, logged, and reconciled regardless of the surface an agent starts from.

FAQ

Common questions from teams building agent workflows.

These are the design boundaries most teams need before wiring assistants into hosting and domain journeys.

Can agents add products to checkout directly?

No. Agents can prepare a recommendation and validate requirements, but final checkout always uses a verified customer handoff.

What can the MCP server do today?

It exposes product listing, domain search, and checkout-requirement tools over JSON-RPC for compatible agent runtimes.

How should agents cite OneNet Servers data?

Use the public pages for narrative answers, the catalog API for structured product fields, and the MCP server for tool-oriented calls.

Next step

Start with the docs, then move into tools.

Read the endpoint contracts, sample requests, and MCP usage patterns before wiring a production agent into checkout-adjacent flows.

Open API docs