​
About MarketAssistant

MarketAssistant is a stock analysis tool built with .NET MAUI, integrated with large AI models. This guide explains how to configure and use SiliconCloud’s AI model services within MarketAssistant.

Powered by SiliconCloud’s robust large language models, this guide will help you:

  • Quickly configure the SiliconCloud API
  • Choose the right AI model
  • Optimize cost and performance
  • Troubleshoot common integration issues

​
Download the Software

Before configuration, ensure you have MarketAssistant installed:

Download Link: MarketAssistant Releases

Download and install the version that matches your operating system.

For a complete feature overview of MarketAssistant, please refer to the official documentation

​
Quick Configuration

​
Step 1: Obtain a SiliconCloud API Key

  1. Visit the SiliconCloud website
  2. After registration, go to the API Key Management Page
  3. Click “Create New Key”
  4. Copy the generated API Key (store it securely)

​
Step 2: Configure in MarketAssistant

  1. Open the MarketAssistant application
  2. Navigate to the Settings page
  3. In the AI Model Configuration section, fill in the following:
    • API Key: Paste your SiliconCloud API Key
    • Model Selection: Choose a suitable model from the dropdown (see recommendations below)
    • Embedding Model: Keep default as BAAI/bge-m3
  4. If stock data is required, obtain a token from ZhiTu website
  5. Click Save

​
Step 3: Verify the Configuration

  1. After saving, try running a stock analysis
  2. If it runs successfully, the setup is correct
  3. If issues occur, check the application logs

​
Model Selection Guide

Use CaseRecommended ModelFeaturesSuitable For
Daily UseQwen/Qwen3-32BFast response, low costIndividual Investors
Professional Analysisdeepseek-ai/DeepSeek-R1Strong reasoning, deep analysisProfessional Analysts
Long Document HandlingTongyi-Zhiwen/QwenLong-L1-32BLong context supportReport Analysis
Complex StrategiesQwen/Qwen3-235B-A22BExceptional comprehensionQuantitative Traders

​
Full Supported Model List

  • Qwen/Qwen3-32B – Qwen 32B version
  • deepseek-ai/DeepSeek-R1 – DeepSeek inference model
  • deepseek-ai/DeepSeek-V3 – DeepSeek V3 version
  • Qwen/Qwen3-235B-A22B – Ultra-large scale model
  • Tongyi-Zhiwen/QwenLong-L1-32B – Long-context language model
  • THUDM/GLM-4-32B-0414 – GLM-4 model by Tsinghua
  • THUDM/GLM-Z1-Rumination-32B-0414 – Deep reasoning model
  • tencent/Hunyuan-A13B-Instruct – Tencent Hunyuan model
  • moonshotai/Kimi-K2-Instruct – Kimi K2 model

​
Best Security Practices

​
API Key Safety

Recommendations:

  • API keys are stored locally only
  • Rotate API keys regularly

​
Data Privacy

  • Stock analysis data is processed locally
  • Only analysis requests are sent to SiliconCloud
  • Vectorized documents are stored in a local SQLite database

​
Optimization Tips

  1. Model Tuning: Adjust model selection based on actual usage results
  2. Batch Processing: Take advantage of batch features like AI-based stock screening
  3. Role Adjustment: Customize analyst roles according to your usage patterns

​
Technical Support

​
Getting Help

  • MarketAssistant Issues: Refer to the Issues section on the project’s GitHub page
  • Community Discussion: Join discussions in relevant technical communities

​
Reporting Issues

If you encounter technical problems, please provide the following information:

  1. MarketAssistant version
  2. The SiliconCloud model you used
  3. Error messages and logs
  4. Steps to reproduce the issue