Use local models like gpt4all

[prenesh0309]

can this be configured to use a locally running gpt like gpt4all. that will reduce the api costs significantly.

[arronKler]

if your model is served as a REST api，you can change the openai call in llm_utils.py by using a http request to your api endpoint instead.

[thebestx121]

can u write a tutorial or some pls

[JAmadoD]

i'd love to find out more about how to do this or where i can find out how to do this I can't find anything and gpt is only giving me so much

[arronKler]

there is some discussion and workaround about this, see: #2158

[lc0rp]

Closing as duplicate of #34