# Integrate with AI

Give your agent the [documentation index](/llms.txt), then choose the product skill below. Each skill covers the customer journey, decisions, security, recovery and acceptance tests—not just API calls.

## Choose a skill

Save the linked file as `.agents/skills/myaza-<product>/SKILL.md` for Codex, or `.claude/skills/myaza-<product>/SKILL.md` for Claude Code, inside your integration repository. Review it before enabling it. These are portable Markdown instructions; they do not install software, request credentials or run commands.

| Product | Integration skill |
| --- | --- |
| Identity verification | [Download skill](/documentation/skills/verification/SKILL.md) |
| Workflows and decisioning | [Download skill](/documentation/skills/workflows/SKILL.md) |
| Business verification | [Download skill](/documentation/skills/business-verification/SKILL.md) |
| Risk Intelligence | [Download skill](/documentation/skills/risk-intelligence/SKILL.md) |
| Screening | [Download skill](/documentation/skills/screening/SKILL.md) |
| Events and continuous monitoring | [Download skill](/documentation/skills/monitoring/SKILL.md) |
| Transactions and fraud decisions | [Download skill](/documentation/skills/transactions/SKILL.md) |
| Travel Rule | [Download skill](/documentation/skills/travel-rule/SKILL.md) |
| Crypto operations | [Download skill](/documentation/skills/crypto-operations/SKILL.md) |
| Address Intelligence | [Download skill](/documentation/skills/address-intelligence/SKILL.md) |
| Investigations and reporting | [Download skill](/documentation/skills/investigations/SKILL.md) |
| Webhooks and operations | [Download skill](/documentation/skills/webhooks/SKILL.md) |

## Give the agent a useful brief

Describe your customer journey, target markets, review team and what your application should do when a result is approved, held or blocked. Share synthetic examples only.

```text
Integrate Myaza Trust for our customer onboarding and crypto transfers.
Start in Sandbox. Read the Myaza product skills and current documentation.
Map our customer reference to externalUserId.
Propose the decision and review flow before changing our policies.
Keep credentials on our server. Verify and deduplicate signed webhooks.
Add executable tests for retries, missing information and blocked outcomes.
Return the implementation, test evidence and remaining Live prerequisites.
```

## Find the exact contract

| Resource | Purpose |
| --- | --- |
| [llms.txt](/llms.txt) | Concise discovery and product links |
| [llm.txt](/llm.txt) | Alias for clients using the singular name |
| [Machine index](/documentation/index.json) | Every guide, product skill and content digest |
| [Complete Markdown](/llms-full.txt) | All published guides, without page navigation |
| [Public v1 OpenAPI](https://trust.myaza.app/api/openapi.json) | Machine-readable public v1 operations and schemas |

Every guide has a **Markdown** link and an `/markdown` URL. Fetch individual guides first; use the full bundle when your tool cannot follow links. The index's SHA-256 values identify the exact content reviewed.

### API families

Do not replace one API prefix with another. They are separate contracts.

| Family | Read first | Boundary |
| --- | --- | --- |
| `/api/kyc` | [Authentication](https://trust.myaza.co/documentation/authentication/markdown), [verification quickstart](https://trust.myaza.co/documentation/quickstart/markdown) | Capture, verification and SDK operations; some are publishable-key safe, others are server-only |
| `/api/v1` | [Risk Intelligence quickstart](https://trust.myaza.co/documentation/risk-intelligence-quickstart/markdown), OpenAPI | Server-side risk, transactions and crypto operations |
| `/api/identity` | [Events](https://trust.myaza.co/documentation/monitoring-events/markdown) | The documented Identity event contract, not a replacement transaction endpoint |
| Dashboard and platform routes | Dashboard UI | Session/role protected administration; not a substitute for customer secret-key APIs |

The v1 schema does not describe every KYC or Identity operation. Read the relevant guide as well as the schema. Follow each endpoint's authentication contract; never expose a secret key in a browser or mobile app.

## Build the whole customer flow

1. **Identify:** use stable customer and operation references, scoped to your organisation and environment.
2. **Collect:** request only the evidence needed for the selected journey.
3. **Decide:** distinguish capture completion, processing, review, policy outcome and execution authority.
4. **Enforce:** your application applies the current decision. Myaza does not operate your wallet or freeze customer accounts.
5. **Recover:** verify webhooks, deduplicate, reconcile uncertain requests and handle out-of-order events.
6. **Audit:** retain the reason, policy version, decision and your application's action without copying sensitive payloads into logs.

For Travel Rule, complete information, a successful connection test and a low fraud score are not interchangeable with transfer clearance. See [wallet authorisation](https://trust.myaza.co/documentation/travel-rule-execution/markdown).

## Before Live

Test permission denial, tenant/environment isolation, duplicate requests, timeouts, missing evidence, rejected decisions and webhook recovery. Use [Sandbox scenarios](https://trust.myaza.co/documentation/sandbox-testing/markdown) and the product testing guide. Ordinary Sandbox traffic may use configured pricing; do not assume all testing is free.

Keep proposed policy changes separate from publication. Your authorised team approves risk appetite, country obligations and Live activation. Report local tests, Sandbox integration, real partner communication and Live acceptance separately.

## Read-only MCP

The Core repository provides an optional `pnpm mcp:integration` process with documentation, v1 schema and example tools. It does not accept customer keys or execute transactions. This is a local integration aid, not a publicly hosted MCP endpoint.