LM Studio Provider

[kyoshiroxmibu]

I didn't see LM Studio being supported provider only Ollama, are there any planning to add LM Studio as a provider

[hydropix]

Hi @kyoshiroxmibu,

Good news, this already works today. LM Studio exposes an OpenAI-compatible server, so you don't need a dedicated provider for it.

Just start LM Studio's local server and select the OpenAI provider, then point the API endpoint at LM Studio:

• Endpoint: http://localhost:1234/v1
• Model: the name shown in LM Studio (e.g. qwen2.5-14b-instruct)
• API key: anything non-empty (LM Studio ignores it)

From the CLI it's the same idea:

python translate.py --provider openai \
  --api_endpoint http://localhost:1234/v1 \
  --model your-loaded-model ...

The same approach works for any OpenAI-compatible backend (llama.cpp, vLLM, etc.). Let me know if you hit any issue.