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

# Inbox

> The Inbox is where emails from energy retailers arrive, get matched to an account, and feed the invoice pipeline — triage, classify, and reprocess email-sourced invoices from one place.

The **Inbox** is where emails sent by energy retailers land. Each account has its own email address; messages addressed to it are captured, matched to the account, classified, and fed into the [UtiliRead](/utiliread) OCR pipeline so their PDF attachments become [invoices](/invoices). Open it from **Inbox** in the **General** section of the sidebar.

<Frame>
  <img src="https://mintcdn.com/utilified/jAVFMAGPv86NPuGP/images/navigation/inbox.png?fit=max&auto=format&n=jAVFMAGPv86NPuGP&q=85&s=54615a75a6e9e0de75b1b9a7fd791af2" alt="UMS Inbox showing incoming email-sourced items" width="1440" height="900" data-path="images/navigation/inbox.png" />
</Frame>

## From email to invoice

<Steps>
  <Step title="An email arrives">
    A retailer sends an email to an account's inbox address. UMS captures it and matches it to the account by the recipient. If no account matches, it is marked `Unmatched — No Account`.
  </Step>

  <Step title="It is classified">
    The message is tagged with an `email_type` such as `Invoice`, `Statement`, `Overdue Notice`, or `Data File`. Non-invoice categories are marked `Skipped — Non-Invoice Category`.
  </Step>

  <Step title="UtiliRead reads the attachments">
    For invoice emails, [UtiliRead](/utiliread) runs OCR over the PDF attachments to extract line items, totals, and meter data.
  </Step>

  <Step title="An invoice is created">
    A matched, extracted attachment becomes an [Invoice](/invoices) record, ready for [Validation](/validation).
  </Step>
</Steps>

## The inbox list

The Inbox is a split-pane review surface: a filterable list on the left, the selected message on the right. Status **filter tabs** group items for triage — **Needs Review**, **Skipped**, **Duplicates**, **In Progress**, **Processed**, **Archived**, and **All** — each with a live count. Above the list, filter by **Account**, **Type**, or a **Search** across subject and sender.

Each row shows the **subject** (bold when unread), **sender**, the received **date**, an attachment icon, and chips for the message's **type**, **status**, and pipeline outcome (for example, `2 invoices`, `OCR failed`, or `OCR review`). Tick the checkboxes to **Exclude** or **Archive** items in bulk. The list auto-refreshes every 30 seconds, and **Load more** pages through older messages.

Keyboard shortcuts speed up triage: `j`/`k` move through the list, `a` archives, `x` excludes, `r` reprocesses, and `Enter` marks an item read.

## Message statuses

A message's `status` reflects where it sits in the pipeline:

| Status                           | Meaning                                                      |
| -------------------------------- | ------------------------------------------------------------ |
| `Received`                       | Captured and awaiting processing                             |
| `Processing`                     | UtiliRead OCR is running                                     |
| `Processed Successfully`         | An invoice was created from the email                        |
| `Processed with Errors`          | Processed, but with issues to review                         |
| `No Data Match`                  | Couldn't match the content to a connection or retail account |
| `Unmatched — No Account`         | No account matched the recipient address                     |
| `Skipped — Non-Invoice Category` | Classifier tagged it as a non-invoice type                   |
| `Duplicate`                      | Matches an invoice already in UMS                            |
| `Failed`                         | Processing failed                                            |
| `Archived` / `Excluded`          | Triaged out of the active queue                              |

## Message detail

<Frame>
  <img src="https://mintcdn.com/utilified/Y-bITl2YbF0gRNFt/images/inbox/inbox-detail.png?fit=max&auto=format&n=Y-bITl2YbF0gRNFt&q=85&s=96ba08bda273e0b3cce0ca21b15baef2" alt="The Inbox split view — message list on the left, selected message detail on the right" width="1440" height="900" data-path="images/inbox/inbox-detail.png" />
</Frame>

Selecting a message opens it as a single card: the **From**, **To**, **Cc**, **Date**, and a linked **Account** header, the rendered email body, and an attachments strip. Three chips below the header are editable in place — click to change the **classification** (`email_type`) or **status**, or toggle **Read**/**Unread**. The context rail shows an **AI Summary** of the message and the OCR result, including any duplicates detected and invoices created.

Header actions let you **Download all attachments**, **Forward** the email to yourself, or **Reprocess with OCR** to re-run the pipeline. The three-dots menu holds **Archive** and **Delete**.

<Info>
  The same inbox appears as an **Inbox** tab on an account's detail page (scoped to that account), and as the **Inbox review** tab in [Validation](/validation) for triaging email-sourced invoices across your portfolio.
</Info>
