With the next MAXQDA release, AI Assist can be connected to locally hosted large language models as an alternative to the MAXQDA Cloud. These models run on your own computer or within your institution\'s network, so no internet connection is required for data analysis.

You can then use AI Assist functions such as summarizing, chatting, and suggesting or applying codes with self-hosted language models of your choice. This opens up new possibilities for projects in which data protection requirements or institutional policies rule out cloud-based services.

In this webinar, we will cover:

• Setting up the connection: How to connect MAXQDA to a local or network-hosted model
• Model selection and hosting: What to consider regarding hardware requirements, memory, and network setup
• Output quality: Why results from local models are currently considerably weaker than those from cloud-based models, and what realistic expectations look like