Backend says Invalid URL on making any request.

[asifrahaman13]

Hi. I tried to install and run the application. I am following the CONTRIBUTING.md page.

I am running in development mode. Both backend and front end is running. But when I try to make any request the backend crashes.

Steps to reporoduce:

Clone the repository.

Follow the CONTRIBUTING.md instructions

[since1q80]

I think i have settled the problem down.When visiting the localhost:3000,just to config in the setting page and everything is fine then.

[asifrahaman13]

@since1q80

Can you explain a bit more with some screenshots.

[since1q80]

You should config the model in the web page via the "setting" button which is at the left buttom corner.

@since1q80

Can you explain a bit more with some screenshots.

[asifrahaman13]

@since1q80

Thanks man. It worked.

I need another help.

How to configure for the SEARXNG. It runs on port 8888

I am getting the following errors when the agent is trying for google search.

No endpoint "http://localhost:32768" is present from searchxng.

So I tried with the 8888. But I am getting the 403 error.

Note that the application is working fine if serch is not requried (where the open ai can give answer by itself)

[Pallavidatadivr]

@asifrahaman13 Hi, Am stuck exactly at this step but using Groq. Did you find any workaround for this?

[asifrahaman13]

@PallaviDaDivr

Nothing is clearly written in the documentation about that. So I am not really sure how to exactly set up. If you find something please let me know too.

[since1q80]

@since1q80

Thanks man. It worked.

I need another help.

How to configure for the SEARXNG. It runs on port 8888

I am getting the following errors when the agent is trying for google search.

No endpoint "http://localhost:32768" is present from searchxng.

So I tried with the 8888. But I am getting the 403 error.

Note that the application is working fine if serch is not requried (where the open ai can give answer by itself)

Did you use the docker which will make the things much simpler and easier?

[asifrahaman13]

@since1q80

That not feasible option for development. :(

[since1q80]

@since1q80

That not feasible option for development. :(

I just tried in development mode and there are some tricky steps which is ok for me. 1.Since the SearNXG is just a service and is started via docker,you could start it.But bear in mind,you do not use the command like "docker run -d --name searxng -p 4000:8080 searxng/searxng" in the terminal to start.But you should use the docker client to start the searxng service like the pic 2.In the config.toml,you should set the port as 4000 [API_ENDPOINTS] OLLAMA = "" SEARXNG = "http://localhost:4000"

[grrowl]

really should update the default config.toml here to match the custom docker-compose

[ItzCrazyKns]

@since1q80

Thanks man. It worked.

I need another help.

How to configure for the SEARXNG. It runs on port 8888

I am getting the following errors when the agent is trying for google search.

No endpoint "http://localhost:32768" is present from searchxng.

So I tried with the 8888. But I am getting the 403 error.

Note that the application is working fine if serch is not requried (where the open ai can give answer by itself)

Hi, you need to add JSON format in SearXNG settings. The SearXNG container we provide already has this settings enabled so you can either use that or add the JSON format in the settings file of your hosted SearXNG.