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

# Trust Accounting overview

> Learn what the Stax Trust Accounting API does and how to start integrating it into your product.

## What is the Trust Accounting API?

The **Stax Trust Accounting (TPA) API** takes in **scanned brokerage statements**, parses them into
structured data, and produces everything you need to generate a **trust report spreadsheet**.

Instead of manually keying in statement data or reconciling accounts by hand, you:

* Upload brokerage statements as documents.
* Attach each document to the correct **plan** and **brokerage account**.
* Let Stax parse the statement into **transactions**, balances, and summaries.
* Use the resulting data to generate trust reports and spreadsheets for your plans.

> The API is designed for **external developers** who want to automate trust reporting based on
> brokerage statements.

## Core concepts

At a high level, the Trust Accounting domain is built around:

* **TPA plans (`TpaPlan`)**: Retirement or trust plans you administer (for example, a specific 401(k) plan).
* **TPA accounts (`TpaAccount`)**: Individual brokerage accounts within a plan, usually identified by
  provider and account number.
* **Documents**: Uploaded brokerage statement files (often PDFs) that you send to the API.
* **TPA statements (`TpaStatement`)**: Parsed representations of a single brokerage statement, including
  beginning/ending balances, the covered period, and transaction summaries.
* **TPA transactions (`TpaTransaction`)**: Line items parsed from the statement (for example, deposits,
  withdrawals, income, fees, transfers, purchases).

You can see the full schemas for these objects in the **API Reference → Trust Accounting API** tab
(`TpaPlan`, `TpaAccount`, `TpaStatement`, `TpaTransaction`, `TransactionSummary`, and related types).

## How the API fits into your architecture

The Trust Accounting API is a **JSON over HTTPS** API layered on top of your existing systems:

```mermaid theme={null}
flowchart LR
  source["Scanned brokerage statements"] --> apiLayer["Trust Accounting API"]
  apiLayer --> parsed["Parsed statements & transactions"]
  parsed --> reports["Trust report spreadsheets"]
```

* Your backend or ingestion pipeline uploads scanned brokerage statements using the **Document** endpoints.
* Once parsed, you use the **TPA** endpoints to attach statements to plans/accounts and work with the
  extracted transactions.
* From there, you can build or download **trust report spreadsheets** based on the parsed data.

For a full list of endpoints and schemas, see the **API Reference → Trust Accounting API** tab in the
navigation.

## Authentication at a glance

All Trust Accounting API requests must include:

* `Authorization: Bearer {apiKey}`
* `email: user@example.com`

This combination authenticates the request and ties it to a specific user identity for auditing.
For example:

```http theme={null}
POST /trust/example HTTP/1.1
Host: api.stax.ai
Authorization: Bearer sk_live_xxx
email: user@example.com
Content-Type: application/json
```

You can find more details, including best practices for key management and environment setup, in
the **Authentication** page under Trust Accounting.

## What you can build

Here are some common product patterns you can implement with the Trust Accounting API:

* **Ingestion pipelines** that automatically upload brokerage statements from custodians or file drops.
* **Review tools** for operations teams to review parsed statements, fix issues, and confirm results.
* **Automated trust reports** that generate spreadsheets per plan, account, or period using parsed data.
* **Audit workflows** that surface discrepancies and confidence scores for further investigation.

## Next steps

To get the most out of the Trust Accounting API, continue with:

* **Authentication** – How to authenticate requests and manage API keys.
* **API endpoints** – A tour of the key operations and workflows.
* **Data model** – The entities, relationships, and IDs you will work with.
* **Limits & performance** – Guidance on rate limits, retries, and performance expectations.