Feature: Local LLM support

[darwpm4gs]

a way to use local LLM? let's say from LM studio

[mastx911]

Forget about local LLMs for gaming. They are extremely resource-hungry (heavy on GPU/VRAM), slow, and often "hallucinate" (invent facts) instead of translating accurately. For a smooth experience, I highly recommend using the DeepL (web) method — it’s faster, higher quality, and free:
#112 (comment)

But if you still want to play around with a local setup, here is a guide for you:
#112 (comment)
Note for LM Studio: It works exactly like Ollama. (But slower than Ollama). You just need to change the API URL in the script:
From: http://localhost:11434/v1/chat/completions
To: http://localhost:1234/v1/chat/completions
Also, make sure you enter the exact full name of the model you have loaded in LM Studio.

P.S. Just to be clear, I am not the developer of MORT. I’m just a user who has struggled a lot with local LLMs and spent enough time testing them to say: save your time and nerves, go with DeepL!

[killkimno]

Thank you for your feedback. continuing to work on implementing this method.

[windssykwt]

Forget about local LLMs for gaming. They are extremely resource-hungry (heavy on GPU/VRAM), slow, and often "hallucinate" (invent facts) instead of translating accurately. For a smooth experience, I highly recommend using the DeepL (web) method — it’s faster, higher quality, and free: #112 (comment)

But if you still want to play around with a local setup, here is a guide for you: #112 (comment) Note for LM Studio: It works exactly like Ollama. (But slower than Ollama). You just need to change the API URL in the script: From: http://localhost:11434/v1/chat/completions To: http://localhost:1234/v1/chat/completions Also, make sure you enter the exact full name of the model you have loaded in LM Studio.

P.S. Just to be clear, I am not the developer of MORT. I’m just a user who has struggled a lot with local LLMs and spent enough time testing them to say: save your time and nerves, go with DeepL!

thank you so much, you helping me to use custom LLM