What embedding model should I choose if I use deepseek

[Huzo]

Hello!

I am a beginner in using these APIs and to Obsidian Copilot. I do get "Fatal Error When Indexing" error and I suspect it is because I do not have a verified API key for the chosen Embedding Model. I do use deepseek and only have a deepseek API verified. In this case, does this mean I also have to pay for some other AI model since deepseek is not in the list? I am a bit confused and some help would be much appreciated.

[logancyang]

Hi @Huzo, welcome to the community!

If you are a paid user of Copilot Plus, you can just apply your license key and use one of the embedding models that starts with copilot-plus.

If you are a free user of the Copilot plugin, yes you do need an embedding provider along with the chat provider. DeepSeek is a chat model provider. So you need to set up an embedding model. You have 3 options:

1. Set up an account and get an API key from the list of Embedding Providers such as OpenAI or Gemini, or
2. Add a custom embedding model using a provider of your choice: https://www.obsidiancopilot.com/en/docs/settings#adding-custom-models, or
3. Use a local embedding model via Ollama or LM Studio: https://github.com/logancyang/obsidian-copilot/blob/master/local_copilot.md

I recommend 1 or 2 if you are unfamiliar with setting up a local server.