Perplexica’s supports Configurable private deployment model

[wangyongpenga]

Perplexica’s supports Configurable private deployment model by OpenAI-compatible interface

[Wangwentao1]

you can use custom openai in the settings it is ok

[wangyongpenga]

you can use custom openai in the settings it is ok

How to configure it here if i use custom openai ?

[Wangwentao1]

you can just open the webpage and set the api to custom api mode，Like the picture：

You dont need to modify the config toml ,just launch the perplexica and press the config button

[wangyongpenga]

you can just open the webpage and set the api to custom api mode，Like the picture：

You dont need to modify the config toml ,just launch the perplexica and press the config button

Avatar
I have configured it on the page, but when I call the API to retrieve the model, the configured model is not obtained.

[Wangwentao1]

you need to press the upload button man！

[rqi14]

you can just open the webpage and set the api to custom api mode，Like the picture：

You dont need to modify the config toml ,just launch the perplexica and press the config button

Hi. Whatever model is set, it automatically changes back to undefined after clicking the upload button. Do you have an idea how it might be fixed?

[Wangwentao1]

the model name is provided by your provieder