Frequently Asked Questions (FAQ)

This section compiles frequently asked questions about Agent for LabVIEW during installation, activation, API configuration, example execution, tool calling, and deployment.

I. Installation and Environment

Q: Which LabVIEW versions are supported? A: Supports LabVIEW 2018 and above, both 32-bit and 64-bit (need to install the corresponding bit count toolkit).

Q: Installation error This package is not compatible with any LabVIEW version on this computer该怎么办? A: Usually caused by bit count or version mismatch. Please check:

  • LabVIEW version is >= 2018

  • Installation package bit count matches LabVIEW bit count (32 for 32, 64 for 64)

  • VIPM is version 2021 or above (recommended to run in administrator mode)

Q: Why can't I see the function palette VIRobotics -> AI Agent after installation? A: Please restart LabVIEW and check again; if still not visible, it is recommended to reinstall the toolkit and confirm the installation status is No Errors.

Q: Is Python installation required? A: Not required. Agent core functionality can be used directly in LabVIEW; Python environment is only needed when running additional Python scripts generated by the Agent.

II. License and Activation

Q: Is trial available? A: Yes, trials are supported. Specific duration and policy are subject to the current release notes.

Q: License State shows No license该怎么办? A: In Tools -> VIRobotics License Manager, select the corresponding product (such as genai), click Apply to apply for trial, or execute formal activation.

Q: Online activation failed该怎么办? A: Priority check:

  • Is License ID / Key entered correctly

  • Is Product Name selected correctly

  • Is device network available

Q: Offline activation failed该怎么办? A: Offline authorization files are bound to devices. Please confirm that .lic matches the current machine's Computer ID. If they don't match, you need to reapply.

III. API Key and Model Calling

Q: API calling failed after configuring API Key该怎么办? A: Check in order:

  1. Key is complete and correct (no spaces, no truncation)

  2. Service provider account has balance/calling permissions

  3. Server selected in LLM Keys matches Key source

  4. Network/proxy/firewall is not blocking requests

Q: basic.vi has responses, but stream.vi has no output, how to troubleshoot? A: Usually caused by model's streaming capability or network issues. Please switch to a model confirmed to support streaming first, and check API balance and network stability.

Q: Which model service providers are supported? A: Supports Aliyun, Baidu, DeepSeek, Doubao, Kimi, Ollama, Siliconflow, Zhipu, etc. (subject to actual version).

IV. Examples and Development

Q: Where are the examples? A: Help -> Find Examples -> Directory Structure -> VIRobotics -> AI Agent.

Q: Should I look at examples by provider or by capability? A: It is recommended to learn by capability (basic chat -> streaming -> tool calling -> comprehensive Agent), and the quick index by provider is for "users with only a single API Key" scenario.

Q: When is AI_Agent_Full.vi appropriate to use? A: Suitable for comprehensive verification. It typically covers streaming output, tool calling, and context management, and is the recommended primary practice example.

Q: Can I modify official examples directly? A: Yes, but it is recommended to "Save As" a new file before modifying to avoid overwriting the original examples.

V. Tool Calling

Q: Why doesn't the model call the tool I wrote? A: Common causes are unclear JSON descriptions. Please focus on checking:

  • Is name clear

  • Is description written to clarify "when to use this tool"

  • Are parameters complete and types correct

Q: Tool was called but execution error该怎么办? A: Run VI standalone first to confirm inputs/outputs and dependencies are normal; then integrate into Agent for joint debugging.

Q: Tool executed successfully but model response is abnormal该怎么办? A: Check if tool result return format meets the requirements of Generate_Tool_Call_Results.

VI. Image Understanding and Image Generation

Q: Why do ImageGeneration-related examples fail? A: Please confirm the following prerequisites:

  • Doubao-related model capabilities are activated (including doubao-seedream)

  • Doubao API is configured correctly

  • AI Vision Toolkit for GPU is installed

Q: text_to_img_test.vi can generate images but results are unstable该怎么办? A: Recommendations:

  • Write prompts more specifically (subject, style, constraints)

  • Add negative prompts when needed

  • Keep parameters fixed for comparative testing

Q: Image-to-image or image fusion results are not as expected该怎么办? A: First reduce task complexity (do it in two steps), and improve input image quality.

VII. Deployment and Operation

Q: Does the deployment target need LabVIEW development environment installed? A: Usually not required, but deployment dependencies must be met (such as Runtime, authorization, external dependency files).

Q: After packaging EXE, it reports "unauthorized"该怎么办? A: Please execute the authorization process on the deployment target and confirm the deployment license status is normal.

Q: Normal in development, abnormal in deployment该怎么办? A: Focus on checking:

  • Are dependency files completely packaged

  • Are paths hardcoded for the development machine

  • Are environment variables and runtime libraries complete

VIII. Still Cannot Resolve

Please contact technical support and attach the following information:

  • LabVIEW version and bit count

  • Operating system version

  • Toolkit version number

  • Name of VI with issue

  • Error screenshots and reproduction steps

Contact information: