Connect software factory with local LLM

This indeed will be an excellent feature.

I was wondering ​@Mark Jongeling, do you have an LLM currently learning the whole idea on how the SF works? For example to give advice when or when not to use a template or make it dedicated for a certain item/feature/object, or maybe explicitly suggesting to apply dynamic logic, or to separate/split domains due to their usage in the model etc etc etc...

We are currently working on something more generic than just the SF and IAM. This LLM model would then be able to work effectively for all customers. A demo was given during last years’ Thinkwise Summit. The way you would give information to the LLM via the model would be the same way we could provide info about the IAM and SF models. 

With Thinkwise Platform release 2025.3 we have made it possible to integrate with any OpenAI-compatible AI model. As most LLMs are OpenAI-compatible, this means you can now use almost any LLM with the Thinkwise Platform, including Open-Source models which you could host yourself. Some examples:

• Claude by Anthropic
• Gemini by Google
• Llama by Meta
• Open-Source models available via Ollama

Note that you need at least Indicium 2025.3.12 to get these models working.