# Convert Large Scanned PDF to Word Online: Size Limits

- Author: Schema Law (https://www.schemalaw.com/)
- Published: 2022-12-18T21:31:13Z
- Updated: 2026-10-11T19:22:20Z
- Category: Converting scans
- Canonical: https://www.schemalaw.com/blog/convert-large-scanned-pdf-to-word-online

> A large scanned PDF fails to convert to Word online when it exceeds a megabyte or page cap. Find which limit you hit, then shrink or split the scan.

To convert a large scanned PDF to Word online, find out which limit the file breaks: megabytes or pages. If it is megabytes, save a smaller copy by dropping colour and excess resolution. If it is pages, split the scan into ranges, convert each range, and join the Word files afterwards.

Both problems look the same from your chair: the upload is refused, or it stalls. They have different cures, so the first job is to read two numbers off the file before you change anything.

## How large a scanned PDF can an online converter take?

An online converter sets a ceiling on file size and often a second one on page count, and both vary from service to service. Schema Law accepts a PDF, PNG or JPEG of up to 20 MB on every plan. The page ceiling for a single document depends on the plan you are on.

| Schema Law plan | Largest upload | Most pages in one document |
| --- | --- | --- |
| Free | 20 MB | 3 |
| Solo | 20 MB | 250 |
| Team | 20 MB | 1,000 |
| Enterprise | 20 MB | No fixed page limit |

The figures come from the [pricing page](https://www.schemalaw.com/pricing), which also shows the monthly page allowance for each plan. Word is an export format on the free plan, but that plan allows one document a day, so it suits a three-page test of a hard scan and not a long file. The post on [OCR cost per page](https://www.schemalaw.com/blog/ocr-cost-per-page) works through what the paid allowances come to for a real month of work.

Before you upload anywhere, check the file against both ceilings:

1. **Megabytes.** Right-click the file and open its properties, or look at the size column in the folder view.
2. **Pages.** Open the PDF and read the page count from the toolbar.
3. **Megabytes per page.** Divide the first number by the second. This one figure tells you whether shrinking is realistic.

A 180-page deposition exhibit that weighs 64 MB runs at roughly 0.36 MB a page. At that rate about 55 pages fit under a 20 MB ceiling, so the file needs either a lighter copy or at least four pieces.

## Why is a scanned PDF so many megabytes?

A scanned PDF is large because it stores a photograph of every page. Three scanner settings decide how heavy each photograph is: the colour mode, the resolution, and how hard the image is compressed. A typed page saved from a word processor stores characters instead, which is why it is a tiny fraction of the size.

- **Colour mode.** Full colour records three values for every dot on the page. Greyscale records one. Black and white records a single on-or-off bit and compresses very well. A colour scan of a black-and-white contract carries a great deal of data that says only "this paper is slightly cream".
- **Resolution.** Doubling the dots per inch quadruples the dots on the page. A 600 dpi scan holds four times the pixels of a 300 dpi scan of the same sheet.
- **Compression.** A copier set to store pages with little or no compression makes files that can be many times larger than they need to be.

Blank backs of pages, a scanned folder cover and oversized pages such as plats all add weight too. So does a file assembled from phone photos, which are colour images at camera resolution. If your PDF began life as a set of loose pictures, the post on [turning scanned images into one PDF](https://www.schemalaw.com/blog/scanned-images-to-pdf) covers how to build it sensibly in the first place.

## Shrinking a scanned PDF without hurting recognition

You can usually remove colour and surplus resolution from a scanned PDF without making it harder to read by OCR. What you must not do is compress the page images so hard that the edges of the letters smear. Always shrink a copy, and keep the scan as it arrived.

Safe to reduce:

- **Colour to greyscale** on pages that are text on white or off-white paper.
- **Resolution above about 300 dpi** for ordinary body type. Text at that resolution is already large enough in pixels for recognition.
- **Blank pages,** if nothing depends on the page numbering matching the paper. If it does, leave them in.

Reduce with care, or not at all:

- **Resolution below about 300 dpi.** Small print, footnotes and fax headers break up first.
- **Heavy JPEG compression.** A "smallest file" preset blurs thin strokes and fills the space around letters with speckle. A page that looks acceptable at normal zoom can read badly.
- **Black-and-white conversion of marked-up pages.** Highlighter, coloured stamps and blue ink can turn into solid black blocks over the words they sat on.
- **Colour that carries meaning,** such as a redline printed in red and blue or a colour-coded exhibit.

The cleanest reduction is a fresh scan. If you still hold the paper, rescan at 300 dpi in greyscale and the file will usually come in far smaller than a 600 dpi colour original. The digitizing guide lists the [scanner settings that make OCR work](https://www.schemalaw.com/blog/how-to-digitize-law-firm-paper-files#scanner-settings-that-make-ocr-work). If you only hold the PDF, use the reduce-size or optimize command in your PDF program on a copy, choose the mildest setting, then zoom to 300% on the smallest type in the document. If the letters have held their shape, upload that copy. If they have gone soft, discard it and split the original instead.

## Splitting a long scan by page range and rejoining the Word files

When a scan is over the page ceiling, or still too heavy after shrinking, split it into page ranges, convert each range separately, and combine the Word files in order. Split at natural breaks such as exhibit tabs or the start of a new document, and never in the middle of a table.

1. **Work out the piece size.** Take the megabytes-per-page figure and leave a margin. For the 64 MB, 180-page exhibit, 45 pages a piece gives four files of about 16 MB each.
2. **Extract each range to its own PDF.** Most PDF programs have an extract or split command. Extracting copies the page images as they are, so nothing is recompressed.
3. **Name the pieces so they sort.** For example `Okafor dep ex 12 p001-045.pdf`, `p046-090.pdf` and so on, with leading zeros.
4. **Confirm the page counts add up.** Four pieces of 45 pages must total 180. A page lost at a boundary is the usual splitting error.
5. **Convert each piece and export Word.** Keep the same page range in each export's name.
6. **Rejoin in Word.** Open the first export, put the cursor at the end, and use Insert, Object, Text from File to add the next file. Repeat in order.
7. **Read across every join.** Check the last lines of one piece against the first lines of the next, with the scan open beside you.

Each piece counts as its own document and must fit the plan's page limit for one document. The free plan's limits of three pages and one document a day rule it out for a split scan of any length. A whole production of oversized files is a different job, and the post on batch OCR explains [how a batch should be split before upload](https://www.schemalaw.com/blog/batch-ocr-discovery-documents#how-should-a-batch-be-split-before-upload).

Ask first whether you need the whole file as Word at all. If you plan to edit a twelve-page agreement that sits inside a 300-page closing binder, extract and convert those twelve pages only.

The Okafor piece is a deposition exhibit. Put the witness name, the exhibit number and the page range in each Word file name. Convert the pages you will cite by page and line, and leave the rest of the transcript as a searchable PDF when you only need to search it. Keep the court reporter certificate with that PDF.

## What does Word do with a converted document of several hundred pages?

Word can open a converted document of several hundred pages without difficulty, because recognized text is small. Microsoft's [operating limits for Word](https://learn.microsoft.com/en-us/previous-versions/troubleshoot/microsoft-365/microsoft-365-apps/word/operating-parameter-limitation) give 512 MB as the largest file Word can open and 32 MB for the document text alone. A long scan of typed pages falls far below both.

The difficulty is working in the file, not opening it. A few habits help:

- **Use the Navigation pane.** If the export carries headings, they give you a clickable outline. Fix the heading levels early so the outline is worth using.
- **Switch to Draft view for editing.** Word skips page layout in that view, and scrolling a long file is faster.
- **Hold off on tracked changes.** Correct recognition errors first with tracking off. Thousands of trivial revisions slow a long file down and bury the edits that matter.
- **Check tables by section.** Long files are where a table split across pages goes wrong unnoticed.
- **Save under a new name at milestones.** A long file takes hours to verify against its scan, and you do not want to repeat them.

Consider leaving the document in parts. An appeal record or a set of medical charts is easier to check and to cite as six files of 150 pages, each matched to one source PDF, than as one file of 900. Join them only if someone needs a single document.

## Try it on a real document

Upload a scanned PDF and export it as Word, a searchable PDF or plain text on the free plan. Your documents are never used to train an AI model. [Start free](https://www.schemalaw.com/api/auth/google)