> For the complete documentation index, see [llms.txt](https://hinkal-team.gitbook.io/hinkal/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://hinkal-team.gitbook.io/hinkal/introduction/platform-components.md).

# Platform Components

### Hinkal consists of three tightly integrated products.

#### Hinkal SDK

A developer-first toolkit that allows PSPs, payment processors, neobanks, OTC desks, and enterprises to embed universal confidential settlement directly into their existing infrastructure.

[The SDK enables:](/hinkal/hinkal-sdk/hinkal-sdk.md)

* Private treasury and settlement flows
* Confidential merchant payouts and payroll
* Removal of wallet history and transaction linkage
* Selective disclosure for compliance via viewing keys

All without changing custody, compliance processes, or routing logic.

#### Hinkal Pay

A lightweight application and SDK that enables confidential settlement using existing public wallets and supported chains (**Ethereum, Solana, Tron, and major EVM networks**).

The only on-chain settlement infrastructure that keeps both the sender and the recipient confidential.\
\
When a transaction is executed through [Hinkal Pay](/hinkal/hinkal-pay/features.md), funds are deposited into a confidential balance within the Hinkal smart contract, controlled by the recipient’s existing wallet.

The settlement is executed confidentially. Sender, recipient, and amount are not exposed on-chain.

Once funds are received on the confidential balance, recipient can:

* Execute private payouts to vendors, affiliates, employees, partners, or treasury accounts\
  **Sender, recipient, and amount remain confidential on-chain.**
* Send funds to a public wallet address when needed\
  **The sender and the original settlement remain confidential.**

Hinkal Pay integrates into existing payment flows without requiring new wallets, custody changes, or chain migration.

#### Hinkal Wallet

A self-custodial, multi-chain wallet with public and confidential accounts in a single interface.

The public account functions as a standard on-chain wallet. Users can buy tokens, swap, bridge, send, and receive assets.

Assets can be shielded into a confidential account.

From the confidential account, users can:

* execute private-to-private transfers (sender, recipient, and amount remain confidential)
* send funds to public wallet addresses (the sender remains confidential)
* receive funds into a confidential balance
* perform swaps and cross-chain transactions within the confidential account

[Hinkal Wallet](/hinkal/hinkal-wallet/features.md) connects to decentralized applications in the same way as standard wallets.

The wallet demonstrates the capabilities of the Hinkal protocol.

#### Hinkal API

A REST API whose server runs inside a secure enclave, letting developers integrate Hinkal's privacy layer directly into their own backend or platform.

The API enables:

* Confidential deposits, withdrawals, transfers, and swaps callable from any backend
* Enclave-attested execution, so the server itself can be verified to run the expected code
* The same privacy primitives used across the rest of the protocol, exposed as REST endpoints

Integrates without a frontend SDK or wallet UI - suited for platforms that want to drive Hinkal privacy entirely from their own infrastructure.

#### Hinkal MCP

A [Model Context Protocol](https://modelcontextprotocol.io/) server that lets AI agents move funds between public and private (shielded) balances through natural-language requests.

The MCP server enables:

* Confidential deposits, withdrawals and transfers driven by an AI agent instead of manual calls
* A custodial wallet created server-side on first use, so the agent acts on the user's behalf while their machine holds only an access key
* The same deposit, withdraw and transfer primitives as the rest of the protocol, exposed as tools any MCP-compatible assistant (e.g. Claude) can call

Connects to any MCP client - Claude Desktop, Claude Code, or a custom host - without requiring the agent to understand proofs, UTXOs, or relaying.
