# How Many Credits Does It Take to Translate a Book? Understanding EPUB and PDF Billing

EPUB is billed by source word count, while PDF is billed by page count. Using examples of a 100,000-word ebook and a 300-page PDF, see exactly how ordinary extraction, refined extraction, and translation add up.

When dealing with books running hundreds of pages or hundreds of thousands of words, estimating costs beforehand helps you pick the right file format and processing method. How extraction and translation costs combine depends entirely on whether you submit a PDF or an EPUB.

In PDF Craft, costs are deducted in credits based on actual usage, with no recurring monthly subscriptions required. Credits convert at USD 1 = 1 credit. New accounts receive a trial allowance, perfect for running small initial sample tests; you can subsequently choose a top-up tier on the [recharge page](https://pdfcraft.ai/recharge/). Each task deducts credits according to the submitted file's size and the chosen workflow.

Billing units differ between PDF and EPUB: EPUB is measured by word count, while PDF is measured by page count.

## EPUB Translation: Billed by Source Word Count Measured by the Service

Because an EPUB is already structured as reflowable text, billing is calculated directly from the **source word count** contained in the original book as measured by the service. On the [recharge page](https://pdfcraft.ai/recharge/), the EPUB translation benchmark is displayed per **100,000 source words** (which includes characters when measuring Chinese source text).

Suppose you have an English EPUB whose source word count is measured by the service at 100,000 words:

* If you translate the entire book into Chinese, the billing multiplier is exactly 1 baseline unit (100,000 / 100,000 = 1).
* If the book measures 50,000 words, the multiplier is 0.5; if it is a 250,000-word volume, the multiplier is 2.5.

Readers often ask: **Is bilingual mode more expensive than target-only mode?**
The rate is identical for both. Whether you choose to keep only the translation or insert translated paragraphs after the original text, the service bills strictly based on the input **source word count**. The translated paragraphs added in bilingual mode are not counted as extra words or billed again.

Keep in mind that you cannot estimate word count from printed sheet pages or the file size in megabytes. An ebook with high-resolution illustrations might have a large file size but very few words, whereas a text-heavy philosophy book might be small in megabytes yet hold a massive word count. Billing is always determined by the service's actual count of source words (which can include body text, headings, notes, and other textual elements).

## PDF Processing: Choosing Extraction Depth, Plus Optional Translation

The logic for PDFs differs from EPUBs. PDFs are measured by **physical page count**, with the pricing page displaying rates per **100 pages**.

When calculating usage for a PDF, **you choose between ordinary extraction or refined extraction, and then add translation fees if translation is enabled.**

Suppose we have a **300-page** PDF book (equal to three 100-page units). The cost breakdown for different processing options looks like this:

| Selected Option                      | Billing Breakdown                               | Credit Calculation for 300 Pages                      |
| ------------------------------------ | ----------------------------------------------- | ----------------------------------------------------- |
| Ordinary extraction (no translation) | Ordinary extraction rate only                   | Ordinary extraction rate × 3                          |
| Refined extraction (no translation)  | Refined extraction rate only                    | Refined extraction rate × 3                           |
| Ordinary extraction + Translation    | Ordinary extraction rate + PDF translation rate | (Ordinary extraction rate + PDF translation rate) × 3 |
| Refined extraction + Translation     | Refined extraction rate + PDF translation rate  | (Refined extraction rate + PDF translation rate) × 3  |

If you only want to convert a scanned book into a readable EPUB or extract its text, you only need to choose either ordinary extraction or refined extraction, without paying for translation.
Only when you enable translation, choose a target language, and translate the PDF into another language will the corresponding PDF translation rate be added on top of your selected extraction tier.

For the differences between the two extraction modes—ordinary extraction does not extract footnotes, whereas refined extraction includes footnote extraction and page refinement—see the detailed comparison in [Converting Scanned PDFs to EPUB](https://pdfcraft.ai/blog/scanned-pdf-to-epub/).

## Practical Tips for Planning Long Document Processing

Choosing the right file, extraction mode, and translation output from the start helps you avoid duplicate processing. If you have multiple editions or file formats of the same book, here is how to weigh your options:

### 1. Pick the Best Source File Before Deciding on the Workflow

If you have both a well-formatted EPUB and a scanned PDF of the same book, submit the EPUB for translation. An EPUB is already structured text, meaning translation skips the layout segmentation and optical recognition steps entirely—avoiding a new OCR step and its potential errors.

### 2. Test a Short Excerpt Before Submitting a Long Document

When facing a manuscript of three or four hundred pages, use a local tool to split out a few pages for a preliminary test. Try to select pages that cover the book's varied layouts: one page of dense body text, one with a table or figure, and one with footnotes at the bottom. Submit this excerpt to check:

* Whether ordinary extraction is sufficient for your reading needs or whether refined extraction (with footnote extraction) is necessary;
* Whether your translation instructions guide names and domain terminology as expected.

A short sample helps you dial in the right extraction mode and translation options. Confirming everything on a test run before submitting the entire book prevents the frustration and expense of having to rerun a task due to mismatched settings.

### 3. Download and Local Archiving

Once a task completes, save the result files to your local drive. Downloaded EPUB files are completely standalone. Transferring them across reading devices or importing them into different reading apps requires no new translation task and consumes no additional PDF Craft credits. If you encounter issues opening files or rendering layouts, refer to [troubleshooting translated EPUBs](https://pdfcraft.ai/blog/translated-epub-troubleshooting/) for guidance.