# DeepSeek AI

{% embed url="<https://youtu.be/91lMxtnJuVc?si=02Zlkahf7uEa7h1L>" %}

## **Fetching the DeepSeek API Key** <a href="#fetching-the-deepseek-api-key" id="fetching-the-deepseek-api-key"></a>

In order to fetch the API key, login into [DeepSeek](https://www.deepseek.com/)

### **Setting Up DeepSeek AI** <a href="#setting-up-deepseek-ai" id="setting-up-deepseek-ai"></a>

Go to **Integrations** -> **Artificial Intelligence**, select **DeepSeek** and insert your **API token**.&#x20;

<figure><img src="/files/8cutGPAPjnLpIXf5ymBf" alt=""><figcaption></figcaption></figure>

Once you’ve integrated, you can access DeepSeek’s features directly through the **Flow Builder** under the **Integrations action node**.

### **Using DeepSeek Inside Flow Builder:** <a href="#using-deepseek-inside-flow-builder" id="using-deepseek-inside-flow-builder"></a>

You can use DeepSeek inside the flow builder using the action node. Make an action node and select the Integrations option. From the menu that appears, click on DeepDeek.

<figure><img src="/files/fc4oorTS67iBPspAnaq9" alt=""><figcaption></figcaption></figure>

### **DeepSeek AI Available Actions:** <a href="#deepseek-ai-available-actions" id="deepseek-ai-available-actions"></a>

DeepSeek’s capabilities are comparable to OpenAI Chat Completion or Claude, making it intuitive for seasoned users.&#x20;

<figure><img src="/files/Vzhayrz9J9AREIRQmPUa" alt=""><figcaption></figcaption></figure>

Let’s break down the main features:

#### **Create Chat Completion** <a href="#create-chat-completion" id="create-chat-completion"></a>

Easily create engaging and dynamic conversations by configuring the Chat Completion action.

System messages can be personalized.

<figure><img src="/files/KgH1WYB2CJzoz1BUfWHu" alt=""><figcaption></figcaption></figure>

#### **Clear Remembered Chat History** <a href="#clear-remembered-chat-history" id="clear-remembered-chat-history"></a>

This action clears the “DeepSeek” system field JSON. The use case is to control the history and monitor the execution time of requests to avoid server timeouts and token overconsumption.

<figure><img src="/files/96uzQ4t78pwhgQGp0fXN" alt=""><figcaption></figcaption></figure>

#### **List Models** <a href="#list-models" id="list-models"></a>

**DeepSeek Chat**: A fast, conversational model ideal for quick user interactions.

**DeepSeek Reasoner**: More detailed and structured, perfect for use cases requiring logical processing (e.g., troubleshooting or technical queries).

<figure><img src="/files/BunK12dSG3SW0bDeq0Dt" alt=""><figcaption></figcaption></figure>

### **Building a Test Chat Flow** <a href="#building-a-test-chat-flow" id="building-a-test-chat-flow"></a>

&#x20;After integrating DeepSeek, here’s a sample setup for testing:

<figure><img src="/files/3XyZTi7xa8esI0pDJFdm" alt=""><figcaption></figcaption></figure>

**Sample System Message:**

| <p>You are Chad. You are to engage with the user using small talk. Follow the below guidelines for responses:<br>Guideline for responses:</p><ul><li>reply in first person</li><li>tone of voice is witty</li><li>use fitting emojis</li><li>you are not allowed to mention you are an Al or assume any other persona than what is described above</li><li>you are not allowed to talk about politics or religious</li></ul> |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

Set **Max Tokens** to 200 and **Temperature** to 0.4 for optimal response quality.

Enable **Remember History** for ongoing conversations.

<figure><img src="/files/0dPWiw6VWTQ2pXq27PAl" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/NJNgKOVDi5vNsXSroq8m" alt=""><figcaption></figcaption></figure>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.rapidbott.com/help-center/artificial-intelligencs/deepseek-ai.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
