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

# MCP connector (AI agents)

> Let AI assistants operate your organization through the built-in MCP server.

The platform ships a built-in [MCP](https://modelcontextprotocol.io) server,
so AI agents and assistants (Claude, Cursor, custom agents) can operate your
organization directly — no integration code required.

```text theme={null}
Endpoint:  POST https://api.hyperprop.com/platform/v1/mcp
Transport: Streamable HTTP (stateless, JSON responses)
Auth:      X-API-Key header, Authorization: Bearer hp_live_..., or OAuth
```

## Claude (web / desktop)

Use **Add custom connector**: paste the endpoint URL and leave the OAuth
Client ID/Secret fields empty — the connector registers itself
automatically. Claude opens a Hyperprop consent page where an organization admin
pastes the organization API key once; access then follows that key's
permissions and ends if the key is revoked. (Under the hood: OAuth 2.1 with
PKCE and dynamic client registration.)

## Cursor / Claude Code

Add to `mcp.json`:

```json theme={null}
{
  "mcpServers": {
    "hyperprop": {
      "url": "https://api.hyperprop.com/platform/v1/mcp",
      "headers": { "X-API-Key": "hp_live_your_key_here" }
    }
  }
}
```

## Available tools

**Read** — `list_trading_plans`, `list_trading_accounts`,
`get_trading_account`, `get_account_audit_log`, `list_traders`, `get_trader`,
`list_purchases`, `get_purchase`, `get_billing_summary`

**Write** — `create_trading_account`, `update_trading_account`,
`record_payout`

Every tool call executes the corresponding REST endpoint with your key, so
organization scoping, permissions, validation, audit logging, and webhooks
apply exactly as documented on each endpoint.

<Tip>
  Read tools work with any key; write tools need `write`/`admin` permission.
  Connect a **read-only key** if you want a strictly read-only agent.
</Tip>

`create_trading_account` accepts an optional `idempotencyKey` so agent
retries can't create duplicates, and `record_payout` implements the
documented payout recipe (append to `metadata.payouts`, adjust
`currentBalance`, audit-logged `reason`).


## Related topics

- [Introduction](/introduction.md)
- [Changelog](/changelog.md)
- [Authentication](/authentication.md)
- [Accept an agreement](/platform-api/user/accept-an-agreement.md)
- [List API request logs](/platform-api/organization/list-api-request-logs.md)