DeepL
AITranslate text and full documents with DeepL neural machine translation, improve writing with DeepL Write, and manage glossaries to keep terminology consistent across languages.
What This Integration Enables
DeepL covers the full translation surface a business needs. Text translation handles individual strings, document translation preserves the formatting of whole files, DeepL Write improves phrasing and tone, and glossaries lock in the correct rendering of your product and domain terms. Language detection means the flow figures out the source language on its own. Translation quality is table stakes; the value is making it operational. A ticket in another language should translate, route, and reach the right agent without a person shuffling files to a translator. An orchestration layer runs that end to end and pulls a bilingual reviewer in only where the stakes justify it, and FlowRunner is built for that layer.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Multilingual Support
Support tickets arrive in many languages. The agent detects each ticket's language, translates it into the support team's working language with the product glossary applied, and files it with the original attached. The agent replies in their language, and the response is translated back for the customer. Language stops being a routing problem, and the support team works in one language.
Document Localization
Marketing and product documents need to ship in several languages. The agent runs document translation on each file so the layout is preserved, applies the brand glossary, and returns formatted, translated documents. What used to be a slow round-trip with an external translator becomes a step in the publishing flow, with a reviewer checking the customer-facing copy before it goes live.
Inbound Message Triage
Messages come in across languages and channels. The agent detects and translates each one, then routes it by content. Routine messages are handled automatically in translation. Anything legally sensitive or tied to a large account is flagged and routed to a bilingual reviewer, so machine translation never stands alone on the messages that carry risk.
Human-in-Loop Highlight
Machine translation is excellent and still not something to trust blindly on a contract clause or a customer commitment. When a translation is legally sensitive or customer-facing on a high-value account, FlowRunner routes it through a [human-in-the-loop](/concepts/human-in-the-loop/) step: the agent pauses, presents the original and the translation side by side, and sends it to a bilingual reviewer via Slack. They confirm or adjust the wording. Routine translation runs automatically; a person checks the ones where a mistranslation would cost something.
Agent Capabilities
14 actionsTranslation
1- Translate Text Translates text into a target language using DeepL's neural machine translation. The source language is auto-detected unless specified.
Writing
1- Improve Text Improves the writing of a text using DeepL Write, fixing grammar, punctuation and phrasing while preserving meaning.
Documents
3- Upload Document Uploads a document (docx, pptx, xlsx, pdf, srt, txt, html, xlf and more) to DeepL for full-document translation with formatting preserved.
- Get Document Status Checks the translation status of a previously uploaded document. Returns the status ('queued', 'translating', 'done' or 'error'), the estimated seconds remaining while translating, the billed character count once done, and an error message if the translation failed.
- Download Translated Document Downloads the translated document once its status is 'done' (check with Get Document Status first), saves it to FlowRunner file storage and returns its URL.
Glossaries
6- Create Glossary Creates a multilingual glossary (v3) that enforces custom term translations during text and document translation.
- List Glossaries Lists all glossaries in the DeepL account with their IDs, names, language-pair dictionaries and entry counts.
- Get Glossary Retrieves the metadata of a glossary by its ID, including its name, language-pair dictionaries, entry counts and creation time.
- Get Glossary Entries Retrieves the term pairs of one language-pair dictionary in a glossary. Returns both the raw TSV entries as provided by DeepL and a convenient parsed object mapping source terms to target terms.
- Edit Glossary Updates a glossary (v3): rename it and/or replace the entries of one language-pair dictionary.
- Delete Glossary Permanently deletes a glossary and all of its language-pair dictionaries by its ID. Returns a deletion confirmation.
Languages
2- List Source Languages Lists all languages DeepL can translate from, with their language codes and names.
- List Target Languages Lists all languages DeepL can translate into, with their language codes, names and whether they support the formality option.
Account
1- Get Usage Retrieves the current billing-period usage of the DeepL account: characters translated so far and the character limit of the plan.
Start building with DeepL
$100 in credits. No card required. Connect in minutes.