Configuring the DeepSeek Large Model via the Copilot Plugin in Obsidian#
- Register an account with a mobile phone number from Silicon Flow and complete real-name verification
- Enter the Silicon Flow Model Plaza and select a model such as deepseek-ai/DeepSeek-V3
- Create an API key
- Go to the settings of the Copilot plugin in Obsidian
- Uncheck all options in the second Model and click Add custom model
- Fill in the model name as deepseek-ai/DeepSeek-V3, which is crucial and must be correct
- Fill in the display name as desired
- Select Open Ai Format for the provider
- Fill in the Base URL as https://api.siliconflow.cn/v1
- Fill in the API KEY with the API key applied for above
- Click Verify to validate
- After verification, click Add Model to add the model
- Return to the first option Basic
- In the General's Default Chat Model, select DeepSeekV3(Open AI Format)
- At this point, the required configuration is complete; just open the command panel in Obsidian notes and type Copilot to call the desired command
- The following configurations are optional
- You can also modify the location of the prompt folder in the root directory
- For example, move the prompt folder under the Copilot plugin
- You need to synchronize the modification in the plugin settings, changing the fourth item Command's Custom Prompts Folder Name value to .obsidian/plugins/copilot/copilot-custom-prompts
- You can also modify the save location of conversation notes
- This modification is in the plugin's Basic option under Default Conversation Folder Name
- For example, change it to 2-Outputs/copilot-conversations
- You can also modify the location of the prompt folder in the root directory