Model catalogue · Jina AI
Jina Embedding
An embedding model expressly designed for the German language, developed by a Berlin vendor. Where the corpus is predominantly German, this design decides between usable and arbitrary hits: formulations of legal and administrative register are matched more reliably; for strongly multilingual corpora the multilingual models are the better choice. It is offered via the Open Telekom Cloud, which operates it in German data centres itself and offers its customers a standard confidentiality agreement under section 203 of the German Criminal Code; no US group sits in the chain on this route. We do not carry it on any of our routes.
In the catalogue because you need to know this as well.
What it is good for
- Expressly designed for German - matching the language profile of a German firm
- German model vendor, German operator with its own section 203 agreement for its customers
- Compact model class with correspondingly low running effort
Where it stops
- Designed for German - for strongly multilingual corpora the multilingual embedding models are the better choice
- Embedding arranges the corpus, it answers no question
- No legal fine-tuning; judging the relevance of hits stays with the practitioner
Status through us
The same model, procured differently, is a different legal position. The status above refers to procurement through us.
Which route carries it, with which instrument and under which condition, we set out in full to the firm before it commissions us - section 203(4) requires that the chain can be checked, not that it is published. The full matrix, with the sources behind every route, can be requested in the German catalogue after confirming your e-mail address, and we go through it in the conversation.
Is Jina Embedding the right fit?
We go through your tasks and the route that carries them - and say plainly where it does not.
Arrange a conversation