---
title: "Build an integration with an agent"
description: "Use MCP to choose an app builder, SDK, and checkout architecture, produce a build brief, and verify integration readiness."
canonical: "https://studio.inttegro.com/inttegro-mcp/build-integrations"
format: inttegro-machine-docs/v1
language: "Python"
sdk_version: "v12.0.0"
reader_profile: optional
reader_manifest: "https://studio.inttegro.com/machine/python/readers/index.json"
reader_identification: "Optional: declare your client only if known and your current-session capabilities; use reader_manifest for supplemental recommendations."
preamble_estimated_tokens: 725
model_token_report: "https://studio.inttegro.com/machine/token-report.json"
preamble_questions: 5
preamble_source: "general"
recommended_mcp_tools: ["design_integration","get_integration_guide","check_integration_readiness"]
estimated_tokens: 2014
token_estimate_method: "utf8_bytes_divided_by_4"
---

## Machine preamble

**Summary:** Use MCP to choose an app builder, SDK, and checkout architecture, produce a build brief, and verify integration readiness. This Python representation uses the selected SDK's native names wherever Studio has a verified mapping.

### Five questions

#### Q1. When should the reader use Build an integration with an agent?

Use this MCP document when an authenticated AI client needs governed Inttegro context, guidance, presentation, or operations.

#### Q2. What is authoritative on this page?

The documented MCP catalog defines available abilities and confirmation boundaries. For live state, the structured tool result is authoritative; presentation output is only a rendering of that result.

#### Q3. What page-specific constraint matters most?

Read operations do not authorize mutations. Actions retain their exact confirmation requirements, and payment credentials or OTP must stay in secure Checkout.

#### Q4. Can Inttegro MCP materially help with this task?

Yes, when the reader has an authenticated MCP connection and the task needs live context, guided planning, or a supported operation. The strongest matches are `design_integration`, `get_integration_guide`, `check_integration_readiness`; their exact roles and boundaries are listed below.

#### Q5. How should the reader verify the answer and continue?

Preserve the structured tool result, distinguish bounded lists from complete analytics, and reconcile asynchronous actions through an authoritative read. Continue with [MCP tool catalog](https://studio.inttegro.com/machine/python/inttegro-mcp/tools.md) and [Accept payment with Inttegro Checkout](https://studio.inttegro.com/machine/python/guides/accept-payment-with-inttegro-checkout.md) when those concerns enter scope.

### Recommended MCP assistance

Use these through an authenticated Inttegro MCP connection. This document remains the contract source.

- [`design_integration`](https://studio.inttegro.com/machine/python/inttegro-mcp/tools.md#design_integration) — read-only integration guidance. Elicits the integration goal, AI app builder, server runtime, SDK preference, and checkout approach, then returns a platform-aware build and test plan without accepting credentials.
- [`get_integration_guide`](https://studio.inttegro.com/machine/python/inttegro-mcp/tools.md#get_integration_guide) — read-only integration guidance. Returns a focused implementation sequence and canonical Studio links for authentication, checkout, orders, catalog, messaging, idempotency, or testing.
- [`check_integration_readiness`](https://studio.inttegro.com/machine/python/inttegro-mcp/tools.md#check_integration_readiness) — read-only integration guidance. Audits platform secret storage, server-runtime compatibility, idempotency, errors, state verification, checkout returns, and messaging consent without receiving source code or credentials.

## Document content

# Build an integration with an agent

An Inttegro-connected agent can guide the architecture conversation, give your coding agent a concrete implementation brief, and check whether the result is ready for an end-to-end test. The workflow is read-only and never asks for API keys, customer records, or source-code uploads.

## Start the guided design

Ask the agent:

### Prompt

> **Copyable prompt:**
>
> Design our Inttegro integration. Take me through the decisions, including the
> app builder, whether to use an official SDK, and whether to use Inttegro
> Checkout. When we are done, give my coding agent exact implementation steps,
> initial API operations, canonical Studio links, and an end-to-end test plan.

The agent calls [`design_integration`](https://studio.inttegro.com/machine/python/inttegro-mcp/design-integration.md). A client that supports MCP form elicitation presents five choices:

1. Primary goal: hosted checkout, custom checkout, orders and invoices, catalog and buy links, customer messaging, or a broader commerce workflow.
2. Trusted server runtime: TypeScript, Go, Python, PHP, Ruby, Java, .NET, or another runtime.
3. SDK preference: official SDK or direct HTTPS.
4. Checkout approach: Inttegro Checkout, custom checkout, or no checkout in the initial workflow.
5. App builder: Base44, Bolt, Google AI Studio, Lovable, Replit, v0, or another
   build environment.

Clients without form elicitation can pass the same typed choices directly. The tool is stateless: all decisions travel in the tool request or the MCP input response, so the workflow behaves consistently across compatible harnesses.

## What the plan contains

The result includes:

- the selected architecture and recommended path;
- the platform's real server and secret-storage boundary;
- an SDK install command when the selected runtime has one;
- the smallest initial set of Inttegro API operations;
- ordered implementation steps;
- explicit idempotency, secret-handling, timeout, retry, and state-verification rules;
- an end-to-end test plan using disposable fixtures;
- canonical Studio documentation links; and
- one self-contained `agent_brief` that can drive a coding task.

The plan does not invent undocumented request fields. Your coding agent should use the linked Studio pages as the contract source and keep `INTTEGRO_API_KEY` in trusted server infrastructure.

## Ask for a focused guide

Use `get_integration_guide` when the architecture is already known and the agent needs one focused sequence:

- `authentication`
- `hosted_checkout`
- `custom_checkout`
- `orders_and_invoices`
- `catalog_and_buy_links`
- `customer_messaging`
- `idempotency`
- `testing`

For example:

> Give my coding agent the hosted checkout guide. Include the authoritative state-verification step and canonical documentation links.

## Check readiness

After implementation, ask:

> Check whether our hosted Inttegro Checkout integration is ready for an end-to-end test. The API key is server-side, idempotency keys and error handling are implemented, order lookup verifies state, and success and cancellation return paths are configured.

`check_integration_readiness` reports each control as `pass`, `fail`, or `not_applicable`, then returns blockers and the exact end-to-end test sequence. It evaluates declared architecture controls; it does not receive secrets or inspect a private repository.

## Important boundary

These tools help an agent design and verify an integration plan. They do not create API keys, deploy code, inspect private source, or prove a production integration by themselves. Run the returned tests against a disposable Inttegro application and verify authoritative resource state before production rollout.

Continue with [Build with AI](https://studio.inttegro.com/machine/python/ai-builders.md), [Authentication](https://studio.inttegro.com/machine/python/api/authentication.md), [SDKs](https://studio.inttegro.com/machine/python/sdks.md), [Accept payment with Inttegro Checkout](https://studio.inttegro.com/machine/python/guides/accept-payment-with-inttegro-checkout.md), or the complete [MCP tool catalog](https://studio.inttegro.com/machine/python/inttegro-mcp/tools.md).