App Version

v3.25.23

API Provider

LM Studio

Model Used

qwen3-coder or any coding LLM's I have tried that will fit on my system

Roo Code Task Links (Optional)

No response

🔁 Steps to Reproduce

Give Roo any long tasks where processing the prompt will take > 5 minutes. This reliably happens when using Orchestrator mode and Roo switches between any modes.

💥 Outcome Summary

Expected LM Studio to finish processing the prompt before being cut off by timeout from Roo.
I know this is a duplicate of #6521 However it is marked as completed. The solution says it added a roo-cline.apiRequestTimeout which is now in the settings. The default of 600s is greater the 5 minutes but my Roo reliable timesout at 5 minutes every time. Setting apiRequestTimeout to 0 which should disable the timeout results in an immediate timeout. I also tried the solution in #6570 which was to change my LMStudio providers to openai compatible and then add "openAiRequestTimeout": 1800000 to the provider in the roo-code-setting.json, export/import back into roo code but there was still to fix for me. I suspect the "openAiRequestTimeout" attribute has not been implemented because when I exported roo-code-setting.json again this update I made was missing.
If this is a know issue that's being worked on then please ignore.
Apologies for not hyperlinking the other issues I have referenced. For the life of me I could not find the docs on how to do it. It's probably right in front of me.