Skip to main content

Running a Batch

Once your conversion is set up, running a batch is straightforward: upload the pages, wait for processing, then review the results.

Start a batch​

From the Conversions page, find your conversion's card and click Run a batch.

Add the Pages​

Upload the material you want to convert — one PDF, or a photo of each page, in the order they should be read. Not both at once.

  • JPEG, PNG, WebP, or TIFF for photos, one file per page; or a single PDF
  • Drop them on the upload area, or click to browse and select several at once
  • Reorder as needed — each page has Move earlier and Move later buttons

The system accepts up to 200 pages in one batch. If you have more, split the register into smaller batches.

Tips:

  • Photos should be clear, level, and in focus
  • Phone cameras work fine (even high-resolution ones)
  • Each photo must be under 50 MB (most phone photos are 3–8 MB)

Once you're happy with the order, click Use these N pages.

Processing Starts​

The system now:

  1. Takes a first look — scans all pages to spot anything unusual before reading (duplicates, blank pages, unexpected layout)
  2. Reads the text from every page (OCR if they're photos; uses existing OCR if you provided transcripts)
  3. Extracts entries — finds each record and fills in your template fields
  4. Checks for issues — flags entries and pages that look uncertain

The batch pauses after the first look to let you remove any duplicate photos or pages that won't work. You'll see a screen showing what the system found, with the chance to confirm or remove pages before it starts OCR.

You can close the tab and come back later. The batch keeps running in the background. You'll see the in-progress batch on your Conversions page and can click it to check progress.

When should these records go into your catalogue? The run screen asks, beside the pages, and the answer is this batch's own:

  • Someone checks these before they go into the catalogue — the records are made and held to one side, and go in when somebody has read them against the page
  • Put them into the catalogue as they're read — every record goes in as soon as its page has been read

It comes pre-set from your workspace's own answer (Settings → Retroconversion), so a workspace that has never touched it sees the choice already made. A volunteer sees it made for them: records they convert are always checked by someone else first.

Progress indicator:

  • Waiting for your go-ahead — the first look is done; the system is waiting for you to confirm the pages before reading starts
  • Being read — the system is processing pages after you've confirmed them
  • Ready to check — extraction is done; entries are waiting for your review
  • Checked — you've started reviewing and accepting entries
  • Committed — the records are in your catalogue
  • Stopped early — the run ended before it finished

Finding Your Batch Again​

Every batch you've run is listed on your Conversions page under Batches, newest first. Click any batch to go back into it, at any point — tomorrow, next week, or months later.

The list shows:

  • The template name and when you started it
  • How many entries are ready to check
  • How many you've already checked
  • What state it's in

You never have to keep a tab open to come back to a batch.

Next Steps​

Once processing finishes and the status shows Ready to check:

You can also review while it's still processing — the interface shows which pages are done and which are still being read.