DeepL Translator: PDFs, Glossaries & Academic Reading

What DeepL Translator does

DeepL Translator serves a different task from DeepL Write: translation across languages rather than same-language revision. For academic reading, the key decisions are which document route you can use, whether the output is editable, how a glossary applies and what the translated wording actually supports. This profile follows official documentation and includes an original teaching exercise; it reports no translation-quality benchmark or completed file trial.

Choose the translation handoff you need

Choose a documented route and inspect its actual output. These starting points do not rank translation quality.

Where it fits, and where it falls short

A suggested DeepL Translator workflow

1. 1

   Keep the permitted original and identify the passage you need to understand.

2. 2

   Choose the language pair, glossary and output route before translation.

3. 3

   Compare the translated wording, qualifications and document layout with the original.

4. 4

   Retain the source locator and record the assistance actually used.

Translate a source or edit its expression in the same language

Choose Translator when a passage needs to move between languages; choose Write when you want to inspect wording within a supported language. Neither task establishes that a source supports your interpretation. Keep the original title, author and page or section beside the translated reading note rather than replacing the source record with a machine-produced passage.

The Word guide distinguishes selected-text translation from whole-document translation. Replacing selected text changes that selection; its documented whole-document route creates a new translated copy and leaves the original file unchanged. These are different actions. Decide which record you need before choosing a replacement control.

• Inspect DeepL Write for a separate same-language editing pass

Supporting documentation: [1] [14]

Check format, size, characters and editability separately

The current file-format table lists an account without a subscription with 5 MB and 100,000-character limits per PDF, Word or PowerPoint file. It separately says translated files without a subscription are not editable. Passing a file-size check does not establish that you can revise the downloaded result.

Other file types and larger limits depend on the plan, and legacy subscriptions have their own table. Confirm the exact row for your account rather than copying an API ceiling into a website workflow. If the task requires an editable translated working document, settle the output entitlement before translating a long reading.

Supporting documentation: [2]

Inspect a translated PDF before relying on its layout

The PDF guide describes scanned documents but notes that scan quality, custom fonts and large images can affect the result. CAD-exported PDFs are unsupported. If an unsatisfactory PDF came from a Word or PowerPoint document, the guide suggests translating that original source file where available.

PDF-to-DOCX output is a specified paid feature, with Team and Business named in both reviewed guides. Do not infer access for every Pro tier. The editing guide explicitly warns that translation errors and different word lengths can shift the layout. Inspect captions, tables, footnotes and reference markers in the actual downloaded file; format preservation is a feature description, not a fidelity result we measured.

Supporting documentation: [3] [4]

Budget target-language outputs, not only uploaded files

The web file-translation workflow allows one or more target languages and deducts one file translation for each selected target. Its desktop instructions describe a single-target route. Check the actual surface instead of assuming every integration has the web controls. In a mixed-source-language batch, manually selecting one source language applies it to every file and can cause errors.

The allowance guide directs subscribers to Subscription or the file-translation toolbar and says unused included file translations do not carry into the next month. Keep an approved output you intend to reuse. The reviewed sources do not establish that translating the same file again is free or that every unsuccessful attempt is refunded.

Supporting documentation: [5] [6]

Choose a term pair without treating it as exact search and replace

A DeepL glossary maps a word or short phrase in the source language to a preferred target-language translation. The overview says entries adapt to the target language’s grammar, so you do not need to supply every plural or inflected verb form. Language variants are specified on the target side. This is different from Write’s controls for keeping a term unchanged.

For a file, select the glossary before starting and make its source and target match the translation. A valid term mapping cannot preserve every surrounding qualification automatically. Inspect the sentence containing the term, especially negation, causal language and the distinction between an observed result and a proposed explanation.

Supporting documentation: [7] [8]

Keep the terminology CSV distinct from the translated reading

The download guide lists glossary CSV export for specified paid tiers through the web and desktop apps, not mobile apps. Its four fields are the source term, target term, source-language label and target-language label. Check the direction before reusing it. A glossary export does not back up the original document, translated passage or your reading notes.

The upload guide describes UTF-8 CSV beginning with the first entry, without a header or comments. A single language pair uses two columns; multiple pairs use four with both language labels. Confirm the detected separator and pair. These import/export controls have their own Required plan headers; they are not automatically included in every subscription. No glossary transfer was tested for this profile.

Supporting documentation: [9] [10]

Check the language and the particular glossary route

The ordinary glossary language list includes Ukrainian. The AI glossary generator has a narrower language list that does not include it and does not allow language variants. The generator extracts candidate terms from previous translations; DeepL recommends high-quality reviewed source material. Do not copy the ordinary glossary’s language coverage into the generator.

Platform labels also need a task-specific check. The glossary overview lists mobile among supported platforms, but the mobile file-translation guide explicitly excludes glossary support for that file route. Choose a documented web or desktop workflow when a document must use your glossary, rather than assuming a phone exposes the same controls.

Supporting documentation: [7] [11] [12]

Distinguish temporary processing from saved terminology

DeepL’s policy describes free inputs and corrections being used temporarily for network improvement or training and says free-service inputs must not contain personal data. For covered Pro functions, it describes temporary processing without using the content to improve the service. Assets you ask to save, such as glossaries and saved translations, have separate storage arrangements. These are provider policy descriptions, not a security audit or permission to upload confidential coursework.

Local glossary storage is not a blanket no-transmission promise. The policy separately includes free-file glossary content in translation metadata retained for 14 days. That period applies to the specified metadata, not all account records. Confirm the covered function and the institution’s approved arrangement before sharing identifying material.

Supporting documentation: [13]

Preserve the unmeasured outcome in a translation

Original teaching text — “The interviews describe participants’ views; they do not measure changes in academic achievement.” An illustrative Spanish rendering is: “Las entrevistas describen las opiniones de los participantes; no miden cambios en el rendimiento académico.” Both were written for this profile, not obtained from DeepL or drawn from an actual study.

Compare that rendering with the hypothetical change “Las entrevistas demuestran mejoras en el rendimiento académico” — the interviews demonstrate improvements in academic achievement. That change removes the limit and adds a positive finding. A correct translation of a preferred technical term would not repair it. With your own permitted source, identify what was measured, what was not and how strongly the author states the conclusion before evaluating a translated passage.

• Practise checking the exact claim against its source
• Record the translation assistance and checks you actually performed

Choose the continuing reading workflow before the plan

DeepL’s plan guide directs readers to the current comparison while noting legacy accounts and feature-specific requirements. Translator, Write and API access answer different needs. Compare your required file output, terminology management and source language before relying on a broad Pro label.

For an initial proposed evaluation, use a permitted short document with a known term, negation, table unit and limitation. Inspect the translated file and retain the original locator. Record corrections and allowance use separately. This is an unrun evaluation, not evidence of a learning gain or a ranked comparison with another translator.

Supporting documentation: [6] [15]

Pricing, free access and plan checks

Check the current Translator plan and each feature’s Required plan header. Current and legacy offers coexist. File translations, text characters, glossary management and PDF-to-DOCX access are separate conditions; a Translator subscription does not itself provide translation API access. Confirm the actual account allowance, renewal and output before subscribing. This profile does not quote a universal regional price or free monthly file quota.

Confirm billing period, renewal terms, limits and regional availability on the provider’s current plan page before paying.

Sources and research notes

This guide synthesizes official product information with our editorial assessment of academic use. Suggested exercises are examples you can repeat; they are not results of tests we performed. We do not infer accuracy from a provider’s marketing claims.

1. DeepL Translator: text and file routes
2. DeepL: current and legacy formats, file limits and editability
3. DeepL: PDF translation and DOCX output
4. DeepL: edit a file translation and inspect layout changes
5. DeepL: web and desktop translation steps and target counting
6. DeepL: monthly file allowance and billing scope
7. DeepL: glossary terms, languages and grammatical adaptation
8. DeepL: apply a matching glossary to a file
9. DeepL: glossary CSV download and platform limits
10. DeepL: glossary CSV upload and field formats
11. DeepL: glossary generator languages and reviewed source material
12. DeepL: mobile file translation and glossary exclusion
13. DeepL: free/Pro content, saved assets and document metadata policy
14. DeepL for Word: selection replacement and a separate whole-file copy
15. DeepL: current plans, legacy access and separate API entitlement

Found a changed feature or a factual error? Send the exact page and an official source to info@ai-essay.com. Read our editorial policy for how we handle corrections.