DeepL can be integrated into Phrase TMS as a third-party machine translation (MT) engine, allowing you to use DeepL's translation capabilities directly within your Phrase TMS translation workflows.
Depending on your current plan, the setup for the integration of DeepL into Phrase might vary.
As of June 30th 2026, Phrase no longer supports DeepL as a native engine. To keep using DeepL within Phrase, you need to obtain your own DeepL API key and configure the DeepL plugin as described below.
Set up DeepL in Phrase TMS
To configure DeepL as a machine translation engine in Phrase TMS:
- Ensure that you have an active DeepL API key.
- In Phrase TMS, navigate to Phrase Language AI in the sidebar
- Open an existing MT profile or create a new one
- Click on Connect MT engines and select DeepL
- For Base engine URL, select a region that matches your subscription:
- DeepL Pro API (EU): https://api.deepl.com/phrase/v1
- DeepL Pro API (US): https://api-us.deepl.com/phrase/v1
- DeepL Pro API (Japan): https://api-jp.deepl.com/phrase/v1
- DeepL API Free: https://api-free.deepl.com/phrase/v1
- Enter your DeepL API key starting with DeepL-Auth-Key
Example: DeepL-Auth-Key xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx - Configure any additional settings, such as formality
- Click Validate connection
- Click Add
You can now apply DeepL translations to projects via the project settings or use them as part of a pre-translation workflow.
Use DeepL in your Phrase TMS workflow
DeepL can be used in two main ways within Phrase TMS:
Raw machine translation — Suitable for use cases where a fast, unreviewed translation is sufficient (e.g., internal communications). MT is generated automatically during pre-translation.
Machine translation post-editing (MTPE) — For higher-quality output, professional linguists review and edit the raw machine translation. Post-editing decisions can be stored in translation memories and term bases for future reuse.