Model Setup (3 Options)
After reading, you can: Choose between “HuiAI Login Token”, “Custom API Key”, or “Local Offline Models (Ollama)”, connect your AI model within 3 minutes, and start your first conversation.
1. Choose Your Starting Path
Section titled “1. Choose Your Starting Path”To begin using AIPC UP, simply choose one of the three paths that fits your setup:
| Path | Best For | Location & Steps |
|---|---|---|
| Path A: HuiAI Login Token | Users who want zero-config instant access and need the complete suite of extension features | Go to: Settings → HuiAI Config Paste your token and click Save. The model channel is created and synchronized automatically. |
| Path B: Custom API Key | Users with their own API keys (DeepSeek, Moonshot, SiliconFlow, OpenAI, Claude, etc.) | Go to: Settings → Models Click “Add Provider”, select your platform, and enter your API Key to access all basic chat capabilities. |
| Path C: Local Offline Models (Ollama) | Users seeking zero cost, offline functionality, and 100% data privacy | Go to: Settings → Models Connect directly to your local Ollama service. Runs entirely on your local hardware for basic chat. |
Path Capability & Feature Comparison
Section titled “Path Capability & Feature Comparison”When configuring a custom API Key or local model, all basic features (AI chat, Q&A, code generation, project workspace management) work completely normally. However, advanced system-level and cross-device extension features require a HuiAI Login Token:
| Feature | Path A: HuiAI Token | Path B: Custom API Key | Path C: Local Models (Ollama) |
|---|---|---|---|
| Basic AI Chat & Q&A | ✅ Full Access (zero-config) | ✅ Full Access (personal key) | ✅ Full Access (local hardware) |
| Code Generation & Review | ✅ Full Access | ✅ Full Access | ✅ Full Access |
| Workspace & File Management | ✅ Full Access | ✅ Full Access | ✅ Full Access |
| Multi-Agent Collaboration | ✅ Full Access | ✅ Full Access | ✅ Full Access |
| 🖥️ Desktop Automation (OS input control) | ✅ Included | ⚠️ Basic chat works; task execution requires HuiAI Token | ⚠️ Basic chat works; task execution requires HuiAI Token |
| 🌐 Remote WebUI Workspace | ✅ Included | ❌ Requires HuiAI Token | ❌ Requires HuiAI Token |
| 📱 Remote Screen Control / Wall | ✅ Included | ❌ Requires HuiAI Token | ❌ Requires HuiAI Token |
| 🎙️ Smart Voice Input & Dictation | ✅ Included | ❌ Requires HuiAI Token | ❌ Requires HuiAI Token |
2. Path A: HuiAI Login Token (Easiest)
Section titled “2. Path A: HuiAI Login Token (Easiest)”If you have a HuiAI service token, setup takes just one step:
-
Open the Config Page
Click the Settings gear icon in the bottom-left corner of the main interface, and select the first item: HuiAI Config. -
Paste & Save Your Token
Paste your token into the “HuiAI Login Token” input box and click Save Token. -
Verify Connectivity
Click Check Token. When the status shows✓ Token is valid, connection normal, your setup is ready. -
Start Chatting Directly
Upon saving, the system automatically synchronizes and configures the default model channel. Return to the chat window and start messaging right away—no manual model configuration required.
3. Path B: Custom API Key (Transparent & Self-Managed)
Section titled “3. Path B: Custom API Key (Transparent & Self-Managed)”If you have your own API Key from public cloud model providers:
-
Navigate to Models
In the Settings sidebar, click Models (under General Settings). -
Add a Provider
Click the Add Provider button in the top right corner and choose “Manual Configuration”. -
Select Platform & Enter Key\
- Select your provider from the dropdown (DeepSeek, Moonshot/Kimi, SiliconFlow, OpenAI, Anthropic/Claude, Zhipu, etc., or select “Custom” for custom OpenAI-compatible endpoints);
- Enter your API Key;
- Select or type your desired model name.
-
Test & Save
Save and test the connection. Once verified, you can select this model in any conversation.
4. Path C: Local Offline Models (Ollama - Zero Cost)
Section titled “4. Path C: Local Offline Models (Ollama - Zero Cost)”If you prefer to operate entirely offline with zero cloud data transmission:
- Run Ollama locally on your machine with your chosen open-source model (e.g. DeepSeek-R1, Qwen2.5).
- In AIPC UP Settings → Models, connect to your local Ollama endpoint (default
http://localhost:11434). - Completely free, uses zero internet bandwidth, and keeps all computations strictly on your local device.
How to Know It Succeeded
Section titled “How to Know It Succeeded”- Path A: Token status displays
✓ Token is valid, connection normal, and sending a test prompt in chat receives an immediate reply. - Path B/C: The added provider passes the connection test, and you can select the model to converse normally.