stepalibre
Alibre Super User
I'll keep it short...
NIVIDIA ChatRTX is a quick and easy way to utilize AI. My AI projects include a lot more than ChatRTX and a dataset used for RAG, this is the simplest setup to get you started.
NVIDIA ChatRTX User Guide
nvidia.custhelp.com
My Testbed for Alibre Design GitHub projects, contain Alibre help and API scripting source and example files. These are experiments and projects I use to design and development public and private solutions.
I won't cover the literal steps to install and configure ChatRTX. There are plenty of videos and content online to get you up and running. Any Alibre Design AI related questions are welcome.
Telling AI to make a working script (Alibre Script IronPython) or program (AlibreX .NET) is not as simple as telling it what you want. You'll need to understand Alibre APIs, .Net, IronPython2.7 and the dataset you give it. As complexity increases, AI's can make obscure errors and mistakes that become harder to catch, getting lost in the chat history, and logic loops. It's often necessary for you to fix it because the AI simply can't find the correct solution or the solution is bad practice.
Note: programming is not the only AI use case for Alibre Design.
Real example:
@gld "Script won't run" post, and my reasoning
I tested ChatRTX and it passed, finding the correct answer:
Dataset used for RAG:
github.com

Now don't be as confident as I was. AI can and will fill in missing data that it think is logical. In this example it referenced Parameters-with-Units.txt.
github.com
But not Units.py:
github.com

C:\Program Files\Alibre Design 28.0.4.28141\Program\Addons\AlibreScript\AlibreScript.chm
UnitTypes.Nanometers, Micrometers and MicroInches? This is a good example of AI hallucination, or simply, AI is wrong. The answer appear to be correct, UnitTypes.Inches is the correct enum. This can cause you to think, the AI is smarter than it actually is.
Let's understand what's happening here: the dataset has example scripts, some include UnitTypes. Units.py has the three enums being used that match AlibreScript.chm. It's possible ChatRTX did read Units.py but didn't reference it. There are control questions and techniques that can help reduce or avoid this type of AI error.
My posts and ramblings here can help understand the beginning efforts I took to get to this point:
Working examples of AI generated Alibre Design code:
adai.donetoolkit.com
adai.donetoolkit.com
Note: I don't use ChatRTX to generate the code in those databases, I have custom programs that generate them using OpenAI's ChatGPT Pro and APIs. I will use other LLMs and software to generate database content but for now only OpenAI is used.
Datasets:
You can use the content on my GitHub or build your own. When using RAG and Gen-AI with datasets, you may need to restructure the data to produce something that is easy to use as input (prompting) to get the correct output (AI generated script, instruction steps or post processing)
APIs, software and websites I use with my Alibre Design work:
I have paid chat and API accounts for all that offer them.
Hugging Face
LM Studio
Jan
Notion AI
Perplexity Pro
Claude Pro
Grok SuperGrok
Mistral LeChat Pro
Gemini Advanced
ChatGPT Pro
IKI.AI
ChatRTX
They each have unique features, and how I use them varies.
Prompt engineering is very important and now is where I spend the most time; designing and developing prompting systems and tools. FYI, I don’t use the terms prompt engineering or prompting in my AI work I prefer the traditional input and output terminology. The usage here is formal.
My top 5s:
en.wikipedia.org
docs.anthropic.com
github.com
This cookbook is awesome and highly recommended: https://cookbook.openai.com/, https://github.com/openai/openai-cookbook
NIVIDIA ChatRTX is a quick and easy way to utilize AI. My AI projects include a lot more than ChatRTX and a dataset used for RAG, this is the simplest setup to get you started.


What Is Retrieval-Augmented Generation aka RAG?
Retrieval-augmented generation (RAG) is a technique for enhancing the accuracy and reliability of generative AI models with facts fetched from external sources.
blogs.nvidia.com
NVIDIA ChatRTX User Guide
NVIDIA Support
My Testbed for Alibre Design GitHub projects, contain Alibre help and API scripting source and example files. These are experiments and projects I use to design and development public and private solutions.
I won't cover the literal steps to install and configure ChatRTX. There are plenty of videos and content online to get you up and running. Any Alibre Design AI related questions are welcome.
Telling AI to make a working script (Alibre Script IronPython) or program (AlibreX .NET) is not as simple as telling it what you want. You'll need to understand Alibre APIs, .Net, IronPython2.7 and the dataset you give it. As complexity increases, AI's can make obscure errors and mistakes that become harder to catch, getting lost in the chat history, and logic loops. It's often necessary for you to fix it because the AI simply can't find the correct solution or the solution is bad practice.
Note: programming is not the only AI use case for Alibre Design.
Real example:
@gld "Script won't run" post, and my reasoning
I tested ChatRTX and it passed, finding the correct answer:
Dataset used for RAG:
GitHub - Testbed-for-Alibre-Design/alibre-script-examples
Contribute to Testbed-for-Alibre-Design/alibre-script-examples development by creating an account on GitHub.

Now don't be as confident as I was. AI can and will fill in missing data that it think is logical. In this example it referenced Parameters-with-Units.txt.
alibre-script-examples/Parameters-with-Units.py at d679e077d57768d5133727092f5a9e1b138df11d · Testbed-for-Alibre-Design/alibre-script-examples
Contribute to Testbed-for-Alibre-Design/alibre-script-examples development by creating an account on GitHub.
But not Units.py:
alibre-script-examples/Units.py at master · Testbed-for-Alibre-Design/alibre-script-examples
Contribute to Testbed-for-Alibre-Design/alibre-script-examples development by creating an account on GitHub.

C:\Program Files\Alibre Design 28.0.4.28141\Program\Addons\AlibreScript\AlibreScript.chm
UnitTypes.Nanometers, Micrometers and MicroInches? This is a good example of AI hallucination, or simply, AI is wrong. The answer appear to be correct, UnitTypes.Inches is the correct enum. This can cause you to think, the AI is smarter than it actually is.
Let's understand what's happening here: the dataset has example scripts, some include UnitTypes. Units.py has the three enums being used that match AlibreScript.chm. It's possible ChatRTX did read Units.py but didn't reference it. There are control questions and techniques that can help reduce or avoid this type of AI error.
My posts and ramblings here can help understand the beginning efforts I took to get to this point:
AlibreScript Code Assistant - Custom GPT
Stay tuned. https://openai.com/blog/introducing-gpts
www.alibre.com
Working examples of AI generated Alibre Design code:

Database | Notion
Made with Notion, the all-in-one connected workspace with publishing capabilities.


Database | Notion
Made with Notion, the all-in-one connected workspace with publishing capabilities.

Note: I don't use ChatRTX to generate the code in those databases, I have custom programs that generate them using OpenAI's ChatGPT Pro and APIs. I will use other LLMs and software to generate database content but for now only OpenAI is used.
Datasets:
You can use the content on my GitHub or build your own. When using RAG and Gen-AI with datasets, you may need to restructure the data to produce something that is easy to use as input (prompting) to get the correct output (AI generated script, instruction steps or post processing)
APIs, software and websites I use with my Alibre Design work:
I have paid chat and API accounts for all that offer them.
Hugging Face
LM Studio
Jan
Notion AI
Perplexity Pro
Claude Pro
Grok SuperGrok
Mistral LeChat Pro
Gemini Advanced
ChatGPT Pro
IKI.AI
ChatRTX
They each have unique features, and how I use them varies.
Prompt engineering is very important and now is where I spend the most time; designing and developing prompting systems and tools. FYI, I don’t use the terms prompt engineering or prompting in my AI work I prefer the traditional input and output terminology. The usage here is formal.
My top 5s:
Prompt engineering - Wikipedia
Prompt engineering overview - Anthropic

GitHub - dair-ai/Prompt-Engineering-Guide: Guides, papers, lecture, notebooks and resources for prompt engineering
Guides, papers, lecture, notebooks and resources for prompt engineering - dair-ai/Prompt-Engineering-Guide
This cookbook is awesome and highly recommended: https://cookbook.openai.com/, https://github.com/openai/openai-cookbook
Last edited: