Hi Jesse Good , I'm assuming you want to create a Translation Provider plugin, right?

As yr aware, Gemini is a general purpose AI model, so you'll need to prompt it with the source text and other context to translate.  Similar to OpenAI provider for Trados Studio, I imagine.

Before I make anything I just wanted to make sure RWS is not already planning on anything.
The below comment from Daniel Brockmann  seemed to indicate that.

community.rws.com/.../184491

Hi Jesse Good , I see, so you're making reference to working with open source LLM that is available from Hugging Face also.  Yes, we are currently working on a plugin integration with Trados Studio that allows the user to run LLM's locally on their machine/ currently in beta phase.  More to follow on this soon, no official relase dates yet.
Note: I'm managing the service here if yr interested in getting involved: https://github.com/RWS/Hugging-Face-Transformers-Service . There will be major changes to this soon, but this is the current state.
 

Hi,

I'm wondering if it is possible to integrate a client's local AI server in Studio, via a plug-in, API or whatever is needed for that?

Do you have any information about that?

Thanks,

Johannes

Hi Johannes Purer , depends what you mean by local AI server.  Can you ellaborate?

Hi Jesse Good  - personally, I would love for the developer community in general and you in particular :) to develop a plug-in for Gemini that makes use a) of the general translation provider architecture and b) of the new AI Assistant companion API that we released with Trados Studio 2024 SR1 in August 2025. The Hugging Face initiative does not include this, and Google is very reluctant to offer the LLM through Microsoft Azure Foundry and such aggregrators - so it would be great to see if this can be developed in the community. This is not to say that eventually we would get there as well, but this will not happen in the short-term. Does this clarify? Thanks, Daniel

PS: we saw 100+ plug-ins created with the NMT generation, and we are looking forward to a similar adoption with third-party LLM models.

Daniel Brockmann 

>Google is very reluctant to offer the LLM through Microsoft Azure Foundry and such aggregrators
Thanks. That answers my question. 

Hi Team,

Is there any update on Gemini integration with Trados? For my language combination, I find it giving better results than Open AI. It would be fantastic if we had the opportunity to use it as our AI translation aid.