Very confused with process for adding Deepseek for API calls

[wuuulll]

I should start by saying I don't have any sort of coding experience or knowledge regarding .yml files, .env files, and so forth. I got the impression after sifting through the documentation to what seemed applicable for me, that there may be an assumption I have some prerequisite knowledge of how to use and apply certain things, despite not having this prerequisite knowledge or experience. I'm generally quite tech savvy and have no problem following instructions most of the time, including copy & pasting some basic code into a terminal or adding info to files... but I'm simply stumped and confused with trying to get Deepseek added to LibreChat.

I just download LibreChat for the first time yesterday and installed Docker according to the guide and instructions. I'm on Fedora 41/Linux.

Upon opening LibreChat, I noticed DeepSeek wasn't in the dropdown menu at the top. So I began looking through the documentation. This is where I've become completely stumped. I see all sorts of things being mentioned in the documentation, but when it comes to putting things together, I haven't had any clear direction regarding what I should be doing. Hopefully someone could assist, which would be greatly appreciated.

So far, here's what I've done (with Docker closed, not running), which I hope is correct so far, but after that, things are unclear.

1. A librechat.yml file did not exist, only a librechat.example.yml, so I created librechat.yml and pasted the "clean example" text into it from here: https://www.librechat.ai/docs/configuration/librechat_yaml/example

2. A docker-compose.override.yml did not exist, only docker-compose.override.yml.example, so I created docker-compose.override.yml and pasted the suggestion on this page beginning with version:3.4...: https://www.librechat.ai/docs/configuration/librechat_yaml/setup

3. According to this page https://www.librechat.ai/docs/configuration/librechat_yaml/ai_endpoints/deepseek -- I copied the code/text listed to my librechat.yml file and inserted my API key there.

edit -- I also came across a similar thread #5422 where it was suggested that in the .env file under ==Endpoints== I add 'custom' to the list. So for now it reads: # ENDPOINTS=openAI,assistants,azureOpenAI,bingAI,google,gptPlugins,anthropic,custom

----- Yet I reopen Docker and open Librechat (I have it on my desktop dock as a web-app to localhost:3080) and there's still nothing in the dropdown menu at the top showing Deepseek.

What did I do wrong and how can I get this setup properly?

After making these changes with docker closed, I opened it and nothing changed.

Thank you.

[wuuulll]

Update: Well, I got it to show up in the dropdown menu after some troubleshooting! I'm not sure if this made a difference but I did 2 things: I added the API key to the .env file. But I think my main issue was that I wrongly assumed that simply closing Docker from its tray icon, then updating the files, then re-opening Docker would account for the changes. But it seems like that wasn't a substitute for running docker-compose down and docker compose up. After a few hiccups in-between running those and re-launching LibreChat, I now see Deepseek in the list. I gave it a Hello to test a response and I received one back. I just hope I didn't inadvertently touch anything that may impact something else at some point.

[davy1337]

Any reason why you are not using it through OpenRouter? I see you have that one already enabled.

[wuuulll]

Honestly, I'm not sure. Thanks for reminding me. It might make more sense now that I think of it since Deepseek R1, o1 or o1 preview are what I'd likely switch between using and probably easier to monitor usage in one spot, deposit funds, etc. And from what I can tell, there doesn't seem to be any markup in cost.