Welcome to MicroROS V2 Robot Car repository

4. Dify Basic Features

Dify Basic Features

Dify Basic Features1. Course Content2. Start the Dify Service3. Basic Usage3.1 Switching the Dify Language3.1 Connecting Model Provider Services3.2 Switching the AI Application Model4. Account Settings5. Development Documentation

1. Course Content

Understand and master the basic operations and functions of Dify.

2. Start the Dify Service

Open the virtual machine and enter the command in the terminal:


sh ~/bringup_dify.sh

Check the IP address of the virtual machine by clicking "Robot Information" on the left side of the desktop.

Enter the vehicle IP address directly in the browser address bar to access the Dify management page. If the device is logging in for the first time, you need to use the account and password. You can select the language in the upper-left corner.

Note

Account: yahboom@163.com
Password: yahboom123
All account passwords, agent applications, and RAG data are stored locally.

The Dify main interface is shown below:

3. Basic Usage

Tip

If you need to use the cloud AI models of model providers, please make sure the virtual machine is connected to the internet.

3.1 Switching the Dify Language

Generally, the Dify page follows the browser language after opening. If you need to switch manually, you can switch the language in the Settings.

3.1 Connecting Model Provider Services

Dify has built-in model interface plugins for different model providers. These plugins are maintained and upgraded by their respective model providers. We can install the plugins of the corresponding model providers to quickly connect to cloud models from different vendors.
Click "Plugins" in the upper-left of the homepage —> "Explore Marketplace" —> "Model" to enter the model interface plugin page.

Here, the installation and configuration of the Qwen plugin is used as an example. Other plugins use the same method; just click "Install".

Then, in the "Model Provider" section of the "Settings" page, fill in the API-KEY applied for from the corresponding platform. If the API-KEY is valid, the "APIKEY" of the corresponding plugin will show a green light.

Tip

For more detailed API-KEY configuration steps and methods to test the API-KEY, refer to 6. Configure API-KEY in the previous chapter.

3.2 Switching the AI Application Model

For already-developed AI agent applications, you can quickly switch between different models to test the effects of different models. Here, we use the core agent microros-v2-zh of MICROROS-V2 as an example. Open the agent application on the homepage.

There are four core AI components: task routing, decision-layer AI, execution-layer AI, and the box-selection layer.

Here, switching the decision-layer AI is used as an example. Open the "Decision-Layer AI" card and use the dropdown in the model selection bar to switch to models from different providers. Here, qwen3.6-plus is selected.

We can also fine-tune the model's reply effect through parameters. Each parameter can be viewed for its detailed effect by hovering the mouse over it.
Taking the temperature parameter as an example:


xxxxxxxxxx
Used to control the degree of randomness and diversity. The temperature value controls how much the probability distribution of each candidate word is smoothed when generating text. A higher temperature value lowers the peaks of the probability distribution, causing more low-probability words to be selected and making the generated results more diverse; a lower temperature value strengthens the peaks of the probability distribution, making high-probability words more likely to be selected and the generated results more deterministic.

Tip

Beginners can generally use the default parameters; no adjustment is needed.

Note that after modifying an AI application, you need to click "Publish" — "Publish Update" to save the changes.

Warning

Please note

For the execution-layer and box-selection-layer models, because they need to process images, only multimodal models can be used (vision models have a special symbol marker as shown in the figure below).
Task routing and decision-layer models have no restrictions.
Task routing can use small-parameter models to improve response speed.

Box-selection layer model:

If the vision variable is not selected, scroll down to select it.

4. Account Settings

Note

The Dify account is stored locally and has no privacy risks. MICROROS-V2 comes with a preset administrator account at the factory. Only refer to this section of the tutorial if you need to modify the account information.

Click the avatar in the upper-right corner —> Account.

The account information is as follows. You can modify the information as needed.

If you need to log out, click the avatar again to log out.

5. Development Documentation

For users with in-depth development needs later, you can view more detailed development documentation by clicking the avatar in the upper-right corner —> "View Help Documentation".

Then the Dify online development documentation page will open. You can select different documentation content to view from the dropdown list on the left.

Dify Introduction

RAG Knowledge Base Deployment