DeepL

Converts a block of text into one of 30+ supported target languages using DeepL's neural machine translation engine.

Sends a Word, PDF, or HTML document to DeepL and returns a fully translated file that preserves the original formatting and layout.

Analyses a text sample and returns the ISO language code, enabling downstream routing decisions without human classification.

Returns DeepL's current catalogue of source and target languages, keeping your UI dropdowns and routing logic in sync with the live API.

Builds a named translation glossary from term pairs you define, ensuring brand names, product terms, and jargon are never mistranslated.

Retrieves the metadata and term pairs for a specific DeepL glossary by its ID, useful for auditing or selecting the right glossary before a translation job.

Removes a DeepL glossary by ID, keeping the account clean when campaigns end or terminology ownership changes.

Returns all glossaries on the account with their IDs, language pairs, and creation dates, enabling the agent to pick the right one programmatically.

Downloads the full term-pair list from a specific glossary as a structured dataset, enabling audits, diffs, and re-import into other systems.

Translates text using a named glossary to enforce approved term substitutions, ensuring brand and legal terminology is never left to the neural model alone.

Returns the current character count consumed and the monthly plan limit, letting the agent gate translation jobs before hitting the quota wall.

Submits a file (Word, PowerPoint, PDF) to DeepL's async document translation pipeline and returns a document ID for status polling.

Polls DeepL for the current status of an async document translation job — queued, translating, or done — so the workflow knows when to fetch the result.

Retrieves the completed translated file from DeepL once the async job is done, returning the file in its original format with formatting preserved.

Breaks a paragraph into individual sentences according to DeepL's language-aware segmentation rules, useful for preparing text before segment-level translation.

Uses DeepL Write to improve the clarity, tone, and formality of existing text in the same language without changing the meaning.

Fires when a text translation request finishes successfully, returning the translated string and detected source language.

Fires when an async document translation job reaches 'done' status, signalling that the translated file is ready to download.

Fires when a translation request returns an error, capturing the error code and input text so the workflow can retry or alert a human.

Fires when the monthly character usage crosses a configurable threshold (e.g. 80% of plan limit), allowing the workflow to throttle or upgrade before hitting the wall.

Fires when a glossary's term pairs are modified, letting downstream workflows re-validate translated content that relied on the previous version.

Fires when DeepL adds a new source-to-target language pair to its catalogue, enabling the agent to automatically expand available locale options in your product.

Fires when per-minute translation requests exceed a configurable threshold, indicating an unexpected volume surge that may signal runaway automation or abuse.