> ## Documentation Index
> Fetch the complete documentation index at: https://docs.link-token.farhr.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Configure link-token in DeepSeek Harness

[DeepSeek Harness](https://www.deepseek.com/harness/) is an AI coding tool from DeepSeek. You can add link-token as a custom model provider and use link-token-supported models in Harness.

## Prerequisites

Before you begin, make sure you have:

* Installed Node.js and npm.
* Created an API key in the [link-token console](https://link-token.cn/console/token).
* Copied the model ID you want to use from the [link-token model marketplace](https://link-token.cn/pricing).

## Install and start DeepSeek Harness

Run the following command in your terminal:

```bash theme={null}
npx @deepseek-ai/dsh web
```

After the command starts, open DeepSeek Harness in your browser.

## Configure link-token

### 1. Open Settings

Click **Settings** in the lower-left corner.

<img src="https://mintcdn.com/linktoken/CmeMISf2O4oa9WLK/images/DeepseekHarness/en/1.png?fit=max&auto=format&n=CmeMISf2O4oa9WLK&q=85&s=0bc59dd739901502b6d7cf44214460ba" alt="Settings entry on the DeepSeek Harness home page" width="2392" height="1574" data-path="images/DeepseekHarness/en/1.png" />

### 2. Add a custom provider

Open **Models**, then click **Add a custom provider**.

<img src="https://mintcdn.com/linktoken/CmeMISf2O4oa9WLK/images/DeepseekHarness/en/2.png?fit=max&auto=format&n=CmeMISf2O4oa9WLK&q=85&s=cee1319d85239eaa6dda08b805a99334" alt="Add a custom provider button in DeepSeek Harness model settings" width="1776" height="1072" data-path="images/DeepseekHarness/en/2.png" />

### 3. Enter the provider details

Complete the provider configuration with these values:

* **Provider ID**: Enter `link-token`. The value must begin with a lowercase letter.
* **Display name**: Enter `link-token`, or choose another recognizable name.
* **Base URL**: Enter `https://link-token.cn/v1`.
* **API protocol**: Select `openai-completions`.
* **API key**: Paste your link-token API key.

Add the models you want to use under **Models**:

* In the first field, enter the model ID copied from the link-token model marketplace, such as `deepseek-v4-flash-0731`.
* In the second field, enter the name that Harness should display, such as `DeepSeek-V4-Flash`.
* To configure more models, click **Add model** and enter each additional model.

Review the configuration, then click **Apply**.

<img src="https://mintcdn.com/linktoken/CmeMISf2O4oa9WLK/images/DeepseekHarness/en/3.png?fit=max&auto=format&n=CmeMISf2O4oa9WLK&q=85&s=073e38cab93aa4b75a505240c9fff3a2" alt="link-token custom provider configuration in DeepSeek Harness" width="1660" height="1546" data-path="images/DeepseekHarness/en/3.png" />

## Select a workspace and model

Return to the home page and select an existing workspace from the workspace menu. To use a different local directory, click **Add workspace**.

<img src="https://mintcdn.com/linktoken/CmeMISf2O4oa9WLK/images/DeepseekHarness/en/4.png?fit=max&auto=format&n=CmeMISf2O4oa9WLK&q=85&s=0ef09e3b421cdc0baab94c94317e3686" alt="Selecting a workspace in DeepSeek Harness" width="1822" height="922" data-path="images/DeepseekHarness/en/4.png" />

Open the model menu in the lower-right corner of the prompt box. Select the model you configured and the effort level you want. Enter a task and send it to start using the model.

<img src="https://mintcdn.com/linktoken/CmeMISf2O4oa9WLK/images/DeepseekHarness/en/5.png?fit=max&auto=format&n=CmeMISf2O4oa9WLK&q=85&s=f51c6fcfe0de669920d75099b80b54d7" alt="Selecting a link-token model in DeepSeek Harness" width="1688" height="688" data-path="images/DeepseekHarness/en/5.png" />

<Note>
  Never expose your API key in screenshots, logs, or public files.
</Note>