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

# Existing imports

> Monitor background imports, read their results, resume what was interrupted.

Imports run in the background and outlive the wizard. **Backfill → Existing imports** is where you check on them.

```bash theme={null}
npx probe-research            # → Existing imports
probe wizard --action imports
```

Status updates live while the page is open.

## Reading a row

Each saved import shows a mark, a state, a progress bar and a count.

| Mark | State | Means |
| - | - | - |
| `·` | **Queued** | Accepted, not started |
| `»` | **Scanning** | Reading the source, before delivery |
| `»` | **Importing** | Delivering |
| `·` | **Waiting for connection** | The network went away; it will retry itself |
| `✔` | **Complete** | Everything planned arrived |
| `!` | **Partially complete** | Finished, but fewer items than planned — open it to see which |
| `!` | **Interrupted** | Stopped mid-flight; resumable from its checkpoint |
| `✗` | **Needs attention** | Failed; the reason is on the detail |
| `·` | **Canceled** | You stopped it |

> [!NOTE]
>
> **Partially complete is not a synonym for Complete.** A byte-limited import reports the count it actually verified, and the wizard shows it as partial rather than rounding up. A missing count reads as `Completion unavailable`, never as success.

While a session import is re-checking what it already uploaded, the status line says so with its progress — `Checking upload status · 3,200/3,857 checked` — rather than leaving a bar that looks hung.

## What you can do to one

| Action | Effect |
| - | - |
| **Open** | The detail: what was planned, what arrived, what needs review |
| **Resume** | Continue from the saved checkpoint |
| **Cancel** | Stop it. Already-delivered work stays delivered |
| **Back** | Return to the menu. Running imports keep running |

Resuming keeps the original destination and the original file and session versions. It does not pick up new or changed files — those go through `--import-changed` so a retry can never quietly widen its own scope.

## Two bars, not one

Session imports and folder imports are tracked separately and always displayed separately, sessions first. They fail independently and complete independently, and a single merged bar would hide one behind the other.

The same pair appears compactly on the wizard's main menu, directly under the device summary.

## From a script

Background jobs are the wizard's lane. From a shell, `probe backfill` is synchronous and shares the same coverage records and delivery receipts — so a file already delivered by a background job is not uploaded twice.

```bash theme={null}
probe backfill /path/to/work --project existing-project
```

> **[How imports are planned](/wizard/import.md)**
>
> What the scanner reads, what needs review, and how resume works.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.