Setting Up LLM Models

Configure Large Language Models (LLMs) to power your agents and chats.

1. On the left sidebar, hover over User Setting and click on Model Providers.

   

2. Select the LLM provider you want to use (currently only Gemini is supported).

3. Enter your API key for that provider in the designated field.

   

4. Click Add Model.

5. To set a default model for chat, embedding, etc., select Set Default Model and choose the correct models.

   

Note

By default, Discover uses Gemini 2.5 Flash for chat with a thinking budget of 10,000 tokens and gemini-01-embedding for the embedding model.

Your LLM is now ready for use in agents and chats.