AionUI is a powerful open-source, cross-platform AI Agent management desktop tool (30k+ GitHub Stars). It provides a visual interface for managing multiple AI agents and LLM models. This guide shows you how to configure an API proxy (relay station) in AionUI to access AI models like GPT, Claude, and Gemini through a unified gateway like LingTrue.
What is an API Proxy/Relay Station? An API proxy (also called a relay station or API gateway) sits between your client application and the upstream AI model providers. It offers benefits like unified billing, smart routing, cost savings, and access to 35+ models through a single API key. LingTrue (www.lingtrue.com) is one such service.
Prerequisites
- AionUI installed — Download from GitHub Releases (supports Windows, macOS, and Linux)
- An API Key from your relay station — Create a token in your LingTrue dashboard
Step 1: Get Your API Key from LingTrue
- Visit www.lingtrue.com/register and create a free account
- Log into the console dashboard
- Navigate to Token Management and click "Create Token"
- Give your token a descriptive name (e.g., "AionUI")
- Click the Copy button to copy your API Key to clipboard
Your API Key will look like: sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Tip: Visit the Model Pricing page to see all available models and their costs before configuring.
Step 2: Open AionUI LLM Configuration
- Launch AionUI on your computer
- Click Settings (gear icon) in the left sidebar or top menu
- Navigate to LLM Configuration (also labeled as "Model Settings" or "大模型设置")
- Click "Add Model" (添加模型)
Step 3: Configure Custom Platform
In the "Add Model" dialog, configure the following settings:
3.1 Select Platform
Choose "Custom" (自定义) from the platform dropdown. This option supports any OpenAI-compatible API endpoint, which is exactly what LingTrue provides.
3.2 Fill in Connection Details
| Field | Value | Notes |
|---|---|---|
| API Key | sk-xxxx (your copied key) |
Paste the token from Step 1 |
| API Base URL | https://www.lingtrue.com |
Do NOT add /v1 suffix for most models |
| Model Name | e.g., gpt-4o |
Enter the exact model identifier |
3.3 Important: Base URL Format
The Base URL format depends on the model type you're configuring:
- For most models (GPT, Claude, Gemini, DeepSeek, etc.):
Use:https://www.lingtrue.com(without/v1) - For Codex models specifically:
Use:https://www.lingtrue.com/v1(with/v1suffix)
Why the difference? Codex uses a slightly different API path structure. If you're unsure, try without /v1 first — it works for the vast majority of models.
Step 4: Choose Your Model
LingTrue supports 35+ models. Here are the most popular ones you can configure in AionUI:
Recommended Models
| Model | Best For | Model ID |
|---|---|---|
| GPT-4o | General purpose, fast and capable | gpt-4o |
| GPT-4o Mini | Cost-efficient everyday tasks | gpt-4o-mini |
| Claude Sonnet 4 | Coding, detailed analysis, writing | claude-sonnet-4-20250514 |
| Claude Opus 4 | Most capable reasoning model | claude-opus-4-20250514 |
| Gemini 2.5 Pro | Long context, multimodal | gemini-2.5-pro |
| DeepSeek V3 | Cost-effective coding & reasoning | deepseek-chat |
For a complete list with real-time pricing, visit the LingTrue Model Pricing page.
Step 5: Save and Test
- Click "Save" to store your configuration
- Return to the main AionUI interface
- Select the model you just configured from the model dropdown
- Send a test message (e.g., "Hello, can you hear me?")
- If you receive a response, the configuration is successful!
Advanced: Multi-Key Rotation
AionUI supports API key rotation for load balancing and redundancy. To configure multiple keys:
- In the API Key field, enter multiple keys separated by commas:
sk-key1,sk-key2,sk-key3AionUI will automatically rotate between keys when encountering rate limits (429) or auth errors (401), with a 90-second cooldown for failed keys.
Advanced: Configure Multiple Models
You can add multiple model configurations to access different AI models through the same relay station:
- Repeat Steps 2-5 for each model you want to add
- Use the same API Key and same Base URL
- Only change the Model Name field for each entry
This way, you can switch between GPT-4o, Claude, Gemini, and other models directly from AionUI's model selector — all powered by a single LingTrue account.
Troubleshooting
| Problem | Solution |
|---|---|
| Model list is empty | Check that your API Key is valid and Base URL is correct |
| Authentication error (401) | Verify your API Key hasn't expired; regenerate one if needed |
| Connection timeout | Check your network connection; ensure www.lingtrue.com is accessible |
| Rate limit error (429) | Wait a moment and retry, or configure multiple keys for rotation |
| Model not responding | Verify the model name is correct (case-sensitive); check account balance |
| Codex models not working | Make sure to use https://www.lingtrue.com/v1 (with /v1) for Codex |
Summary
Configuring an API proxy in AionUI is straightforward:
- Get API Key from www.lingtrue.com
- Add Custom model in AionUI Settings → LLM Configuration
- Fill in API Key + Base URL (
https://www.lingtrue.com) + Model name - Save and test
With this setup, you can access 35+ AI models through AionUI with smart routing, cost optimization, and a unified billing experience. For more guides, check out our Tech Blog or visit the About page to learn more about LingTrue.