Using Local Models

If you already have a computing power server and have completed the local deployment of large models, you only need to complete the relevant configuration in VJSP to directly call and use this local model.

ⓘ Information

Having multiple configuration files allows you to quickly switch between different models and settings without having to reconfigure everything each time you change settings.

Pre-configuration Preparation

Before starting configuration, please confirm the following information:

Information Item	Description	Example
Server Address	IP address or hostname of the internal network server	192.168.1.1 or ai-server.com
Server Port	Port on which the model service listens	e.g., 11434 or 8000
API Path	Full path of the API interface	/v1
API Key	API key set for the deployed model	sk-xxx...

How to Obtain This Information?

1. Ask the Server Administrator: Obtain server address, port, and authentication information

2. Check Server Documentation: Model services typically have configuration documentation

3. Test Connection: Use curl to test if the API is accessible

# Test if the API is accessible (replace with your server address)
curl http://192.168.1.1:11434/v1

How It Works

Each configuration file can independently maintain the following:

• API key and authentication information

• Model selection

• Temperature settings for controlling response randomness

Setting Up in VJSP

Step 1: Open Model Configuration

1. In the IDE, click the VJSP icon
2. Click the "Settings" interface in the top-right corner
3. Select the "Model Configuration" tab to enter the settings page

Note: default is the configuration file pre-configured by VJSP, which is not recommended to modify. If you need to use local models, please create a new configuration file.

Step 2: Create New Configuration

Click the "+" button to create a new configuration file.

Step 3: Fill in Configuration Parameters

Configuration Item	Description	Example Value
Configuration File Name	Custom name for easy identification	Internal-Qwen3.5 or TEST
API Provider	Select OpenAI Compatible	OpenAI Compatible
OpenAI Base URL	Full API address of the local large model server	http://192.168.1.1:11434/v1
API Key	API key required by the server (if no authentication, fill any value but cannot be empty)	sk-xxx...

Step 4: Select Model

After filling in the correct server address and key, the model list will be automatically loaded from the server into the dropdown menu.

• Select the required model from the dropdown menu
• No need to manually enter the model name

Step 5: Advanced Settings (Optional)

Adjust the following settings based on actual requirements:

Setting Item	Description	Recommended Value
Enable R1 Model Parameters	Enable for QWQ/R1 series models	Check when using R1 models
Use Traditional OpenAI API Format	Compatible with older API	As required by server
Enable Streaming	Support streaming responses	✓ Recommended to check
Include Maximum Output Tokens	Limit output length	✓ Recommended to check

Step 6: Save and Test

1. Click the "Save" button
2. Select the newly configured model in the session window
3. Send a test message to verify the connection

Switching Configuration Files

Two quick switching methods are supported:

1. In the settings panel, directly select the configuration file from the dropdown menu.

2. In the chat interface, use the top API configuration dropdown menu to switch.

   

Pinning and Sorting Configuration Files

You can pin frequently used configuration files to improve switching efficiency:

1. Hover over the configuration file to display the pin icon.
2. Click the pin to pin the configuration file.
3. Pinned files appear at the top of the menu, sorted alphabetically.
4. Unpinned files appear below the separator, also sorted alphabetically.
5. Click the pin again to unpin.

This feature makes it easier to navigate between frequently used configuration files, especially when you have many configurations.

Editing and Deleting Configuration Files

• In the settings, select the configuration file to modify any parameters.

• Click the pencil icon to rename the configuration file.

• Click the trash icon to delete the configuration file (cannot delete the last configuration file).

Security Notice

API keys are securely stored in the IDE's encrypted storage (Secret Storage) and are not exposed in plain text.

Related Features

• Supports integration with Custom Modes

• Supports configuring Temperature Parameters separately for each mode

• Can optimize cost management through Rate Limiting and Usage Statistics