search

janhq / jan

Jan is an open source alternative to ChatGPT that runs 100% offline on your computer. Multiple engine support (llama.cpp, TensorRT-LLM)
https://jan.ai/
GNU Affero General Public License v3.0
23.27k stars 1.35k forks source link

bug: Openrouter API broken with new update #2733

Closed lactoseintolerant closed 6 months ago

lactoseintolerant commented 6 months ago

The Openrouter API authentication no longer functions. Following the instructions at https://jan.ai/docs/remote-inference/router doesn't work. Also, there used to be a box beneath the Openrouter models on the main UI where you could insert the API key. That is gone. Entering the API key from the UI under the new settings/OpenAI extension API input doesn't appear to work, either.

To reproduce, try to run any Openrouter model.

Van-QA commented 6 months ago

hi @lactoseintolerant, recent change with the engine settings causing the need to re-input the API key in the engine settings, please check this out: https://github.com/janhq/jan/issues/2736#issuecomment-2058166270

lactoseintolerant commented 6 months ago

Thanks @Van-QA . I already tried inputting the API key in settings (sorry, I didn't explain clearly above). However, I have figured out how to get it to work. I have to update the endpoint URL in settings to https://openrouter.ai/api/v1/chat/completions Previously, I think you just had to set the endpoint in the individual model json file and the endpoint was unchanged in the engine json file.

bogorad commented 6 months ago

Thanks @Van-QA . I already tried inputting the API key in settings (sorry, I didn't explain clearly above). However, I have figured out how to get it to work. I have to update the endpoint URL in settings to https://openrouter.ai/api/v1/chat/completions Previously, I think you just had to set the endpoint in the individual model json file and the endpoint was unchanged in the engine json file.

oh, it's in UI now!

lactoseintolerant commented 6 months ago

For anyone who is confused: Click gear icon in lower left of UI -> Select OpenAI Inference Engine -> Update Chat Completions Endpoint, Update API Key