> ## Documentation Index
> Fetch the complete documentation index at: https://developers.circle.com/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Before reading these docs, install the relevant Circle Skill to get pre-loaded context — skills are LLM-optimized and cover architecture decisions, correct flows, and common pitfalls. In Claude Code: run /plugin marketplace add circlefin/skills then /plugin install circle-skills@circle. With Vercel Skills CLI: run npx skills add circlefin/skills. Available skills: use-usdc https://github.com/circlefin/skills/blob/master/plugins/circle/skills/use-usdc/SKILL.md (balances, transfers, approvals on EVM and Solana); EURC has no skill — see https://developers.circle.com/stablecoins/what-is-eurc.md; bridge-stablecoin https://github.com/circlefin/skills/blob/master/plugins/circle/skills/bridge-stablecoin/SKILL.md (crosschain transfers via CCTP, USDC, EURC, and third-party assets); use-arc https://github.com/circlefin/skills/blob/master/plugins/circle/skills/use-arc/SKILL.md (Arc chain config, contract deployment, USDC bridging — Arc docs: https://docs.arc.io/llms.txt); use-circle-wallets https://github.com/circlefin/skills/blob/master/plugins/circle/skills/use-circle-wallets/SKILL.md (choose wallet type); use-developer-controlled-wallets https://github.com/circlefin/skills/blob/master/plugins/circle/skills/use-developer-controlled-wallets/SKILL.md (custodial wallets for payouts and treasury); use-user-controlled-wallets https://github.com/circlefin/skills/blob/master/plugins/circle/skills/use-user-controlled-wallets/SKILL.md (embedded wallets with Web2 login); use-modular-wallets https://github.com/circlefin/skills/blob/master/plugins/circle/skills/use-modular-wallets/SKILL.md (smart contract wallets with passkeys and ERC-4337); use-gateway https://github.com/circlefin/skills/blob/master/plugins/circle/skills/use-gateway/SKILL.md (unified USDC balance, nanopayments); use-smart-contract-platform https://github.com/circlefin/skills/blob/master/plugins/circle/skills/use-smart-contract-platform/SKILL.md (deploy, interact, monitor contracts). Pair any skill with the Circle MCP server for live SDK signatures, contract addresses, and chain IDs: {"mcpServers":{"circle":{"url":"https://api.circle.com/v1/codegen/mcp"}}}.
> Read the product overview first. Each product has an overview explaining when and why to use it. Read it before quickstarts.
> Pick the right wallet type. Developer-controlled (you hold keys), user-controlled (end user holds keys via PIN/social login), or modular (smart contract wallets with passkeys). See https://developers.circle.com/wallets.md.
> Use CCTP V2 for crosschain transfers. CCTP natively transfers USDC and EURC via burn-and-mint, and supports permissionless wrapping of third-party assets via lock-and-mint. V1 is legacy; always use V2 unless V1 is specified or the chain requires it (Noble, Sui, and Aptos still require V1). Migration guide: https://developers.circle.com/cctp/migration-from-v1-to-v2.md.
> Use Bridge Kit for frontend bridging. Wraps CCTP with a simpler SDK for user-facing apps. Use CCTP directly for backend transfers.
> Gas Station vs Paymaster. Gas Station sponsors gas for Circle Wallet transactions. Paymaster lets users pay gas in USDC. Different use cases — read both overviews.
> Gateway Nanopayments for sub-cent payments. Gasless USDC micropayments down to $0.000001 via x402 and batched settlement. For pay-per-request APIs, AI agent payments, streaming.
> Look up USDC addresses per chain. Never hardcode — use https://developers.circle.com/stablecoins/usdc-contract-addresses.md.
> Prefer SDKs over raw API calls. Node.js and Python SDKs handle auth, retries, and errors.
> API key required. Bearer token in Authorization header. Testnet and mainnet use separate keys and may use different base URLs depending on the product.
> Set up webhooks when available. Most operations are async. Webhooks deliver transaction confirmations and state changes.
> When calling list endpoints, paginate using pageSize and pageAfter until no nextPageAfter cursor is returned—stopping at the first page silently misses records.
> Building an AI agent? Start with the Agent Stack—Circle CLI, agent wallets, and nanopayments built for autonomous use cases: https://developers.circle.com/agent-stack.md.

# Array sections

> Understand how to work with repeatable entity sections like beneficial owners and directors

Some sections hold repeatable entities such as `beneficial_owners`,
`authorized_signers`, or `directors`. These are **array sections**. The schema
marks them as `type: array` with `items` for each entity. Array sections exist
because a business can have any number of these entities—two beneficial owners,
or five.

Each entity gets its own `refId` so you can update it or attach documents to it
on its own. These sections use different read, save, and document rules than
flat ones.

[Conditional logic](/end-user-onboarding/concepts/conditional-logic) can turn
array sections on or off based on application settings.

## How array sections differ from flat sections

| Operation | Flat section | Array section |
| - | - | - |
| Read | Returns a JSON object | Returns a JSON array; each element includes a `refId` (UUID) |
| Save | `PUT` with a JSON object | `PUT` with a JSON array; include `refId` for existing entities, omit for new ones |
| Delete | Not applicable | `DELETE /sections/{sectionName}/{refId}` removes a single entity |
| Documents | Link with `datumName` only | Link with `datumName` and the entity's `refId` |

## Reading array data

Retrieve array section data with the
[Get section](/api-reference/end-user-onboarding/get-section) endpoint
(`GET /v1/onboarding/partner/applications/{applicationId}/sections/{sectionName}`).
The response is a JSON array where each element includes a `refId`:

```json theme={null}
{
  "data": [
    {
      "refId": "550e8400-e29b-41d4-a716-446655440001",
      "firstName": "Jane",
      "lastName": "Doe",
      "ownershipPercentage": "51"
    },
    {
      "refId": "660e8400-e29b-41d4-a716-446655440002",
      "firstName": "John",
      "lastName": "Smith",
      "ownershipPercentage": "49"
    }
  ]
}
```

<Tip>
  By default, values like `ownershipPercentage` return as strings. Add
  `?typedValues=true` to
  [Get section](/api-reference/end-user-onboarding/get-section),
  [List sections](/api-reference/end-user-onboarding/list-sections) (with
  `includeData=true`), or
  [Get all application data](/api-reference/end-user-onboarding/get-all-data) to
  get native JSON types.
</Tip>

## Saving array data

When saving, include `refId` for entities you want to update and omit it for new
entities:

```shell theme={null}
curl --request PUT \
  --url https://api-sandbox.circle.com/v1/onboarding/partner/applications/${APPLICATION_ID}/sections/beneficial_owners \
  --header "Authorization: Bearer ${YOUR_API_KEY}" \
  --header 'Content-Type: application/json' \
  --data '[
    {
      "refId": "550e8400-e29b-41d4-a716-446655440001",
      "firstName": "Jane",
      "lastName": "Doe",
      "ownershipPercentage": 60
    },
    {
      "firstName": "Alex",
      "lastName": "Johnson",
      "ownershipPercentage": 40
    }
  ]'
```

In this example, the first entity is updated (existing `refId`) and the second
is created (no `refId`). Existing entities not included in the array are
preserved, not deleted.

<Warning>
  Always include the entity's `refId` when you update it. Without `refId`, the API
  adds a new entity instead. That can leave the section `incomplete` if the
  duplicate is missing required documents. Delete it with
  `DELETE /v1/onboarding/partner/applications/{applicationId}/sections/{sectionName}/{refId}`.
</Warning>

## Removing entities

Remove a single entity from an array section with the
[Remove entity](/api-reference/end-user-onboarding/remove-entity) endpoint
(`DELETE /v1/onboarding/partner/applications/{applicationId}/sections/{sectionName}/{refId}`).
The response includes the updated section list with recalculated statuses.

## Document association

When uploading a document for a field in an array entity, include the entity's
`refId` so the document is associated with the correct person or organization.
See [Upload documents](/end-user-onboarding/howtos/upload-documents) for the
full upload workflow.

```shell theme={null}
curl --request POST \
  --url https://api-sandbox.circle.com/v1/onboarding/partner/applications/${APPLICATION_ID}/documents \
  --header "Authorization: Bearer ${YOUR_API_KEY}" \
  --header 'X-Idempotency-Key: ${IDEMPOTENCY_KEY}' \
  --form 'fileContent=@passport.pdf' \
  --form 'fileName=passport.pdf' \
  --form 'datumName=passport_document' \
  --form 'refId=550e8400-e29b-41d4-a716-446655440001' \
  --form 'issuedCountry=US'
```

<Warning>
  Uploading a document for an array entity returns a `documentId`, but that alone
  does not fill the entity's document field. Save the array section again with the
  `documentId` in the relevant field. Include the same `refId` so the API updates
  the existing entity, not a new one.
</Warning>