Developer Guide Introduction¶

This section is intended for developers who wish to perform secondary development, engineering integration, and capability extension using AI Agent for LabVIEW. You will progress from "runnable" to "extensible", gradually mastering the methods to transition from single-feature verification to complete engineering deployment.

Chapter Objectives¶

Through this chapter, you will establish the following capabilities:

Understand the development boundaries and typical development workflows of Agent for LabVIEW
Master the migration path from basic examples to engineered applications
Learn to coordinate tools, models, prompts, and context
Master tool calling integration methods in LabVIEW scenarios
Establish maintainable architecture for future extension modules (e.g., VI-specific modules)

Target Audience¶

Developers wishing to integrate LLM capabilities into LabVIEW
System integration engineers needing to integrate AI capabilities into measurement/control and automation workflows
Technical leads responsible for maintaining team-level toolchains and knowledge assets
Researchers building Agent workflows in teaching or experimental projects
Developers wishing to migrate LabVIEW to other platforms

Recommended Learning Path¶

Stage 1: Complete a Runnable Loop¶

Complete installation and license/API configuration first
Run basic examples to confirm chat and model calling pipeline is functional
Run complete examples to verify context, streaming output, and tool calling

Reference: Installation Guide, License Activation and API Configuration

Stage 2: Enter Developer Mode¶

Familiarize yourself with module organization and calling pipelines
Select a small business scenario to establish a minimum viable flow
Package existing VIs as tools and integrate into Agent
Establish reusable prompts, tool descriptions, and error handling conventions

Reference: Module Overview, Quick Start, Examples and Applications

Stage 3: Engineering and Extension¶

Establish team-level directory structure and naming conventions
Add logging, exception handling, and regression verification
Plan deployment topology (development environment/field environment)
Gradually introduce specialized modules (e.g., VI-specific capability modules)

Reference: Deployment and Distribution, Troubleshooting

Development Architecture Recommendations (Aligned with Current Documentation Structure)¶

It is recommended to split Agent for LabVIEW development content into three layers:

Foundation Layer: Model connection, Key management, session and context management
Capability Layer: Tool calling, document parsing, code assistance, business logic orchestration
Scenario Layer: Test automation, device commissioning, experimental workflows, report generation, and other business scenarios

Based on this structure, VI-specific modules can be naturally extended in the future, such as:

VI parsing and description generation (existing functionality)
VI tool standardization encapsulation (VI + JSON description)
Unified registration and governance for enterprise private tool repositories

Modular Development Practice Recommendations¶

Start with "small but complete" single-scenario prototypes, then extend to multi-scenario reuse
Keep each tool with "single responsibility" - avoid having one VI承担 too much logic
Keep tool inputs and outputs stable, prioritizing observability and debuggability
Treat prompts and tool descriptions as configuration assets, not temporary text

Example Program Entry¶

You can open examples via the following path:

Help -> Find Examples -> Directory Structure -> VIRobotics -> AI Agent

Technical Support¶

If you encounter issues during development, please first consult the subsequent sections of this guide or FAQ:

Technical Support Email: support@virobotics.net
Website: https://www.virobotics.net
QQ Technical Group: 664108337