The Argos Translate data index is a collection of parallel phrase datasets for training machine translation models. Most of this data was collected from Opus, screened for quality, and then packaged in a consistent format for programmatic access.
You can use Argos Open Tech Data with Argos Train or LibreTranslate Locomotive to train a custom neural network for foreign language translation.
I only include datasets on this index if they're available under a permissive license that allows commercial use. You can view the license for each dataset individually on Opus.