> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.vapi.ai/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.vapi.ai/_mcp/server.

# DeepSeek

> Connect a DeepSeek API key to Vapi and configure an assistant with a supported DeepSeek model.

DeepSeek is a large language model provider focused on chat and reasoning models. Connect your own DeepSeek account to use its models on a Vapi assistant.

## What is DeepSeek

DeepSeek builds large language models for chat and reasoning use cases.

## Connect DeepSeek

DeepSeek requires your own account. Connect your DeepSeek API key on the [Integrations](/providers/integrations) page before configuring an assistant. Without a connected DeepSeek key, the models below will not work.

## Language model

Set `model.provider` to `deep-seek` and `model.model` to a supported model.

```bash
curl -X PATCH "https://api.vapi.ai/assistant/ASSISTANT_ID" \
  -H "Authorization: Bearer $VAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": { "provider": "deep-seek", "model": "deepseek-chat" }
  }'
```

For additional API configuration options, review the [`DeepSeekModel` fields](/api-reference/assistants/create#request.body.model.DeepSeekModel) in the Create Assistant API reference.

| Model       | Model ID            |
| ----------- | ------------------- |
| DeepSeek-V3 | `deepseek-chat`     |
| DeepSeek-R1 | `deepseek-reasoner` |

## Related

#### [Integrations](/providers/integrations)

Connect your DeepSeek API key.

#### [Language models](/providers/model/overview)

Compare every language model provider Vapi supports.