Basic idea
Your new RTX Spark is amazing at local content creation. It accelerates 3D, Video and 2D workflows. But it does much more! Thanks to its Tensor Cores, unified memory and NVIDIA RTX technology, it can run large AI models and agents to explore creative ideas and help with work in your creative apps.
In this playbook you’ll learn:
- How to use local creative AI models with ComfyUI.
- How to use a local agent as a creative assistant.
Prerequisites
- An RTX Spark with 32 GB or more of unified memory. Your agent’s AI model and ComfyUI share this memory, so on a 32 GB RTX Spark use the smaller image model we recommend in this playbook and keep the local model your agent picked for your device. Larger units can also run the full-size image model and larger agent models at the same time.
- About 30 GB of free disk space for ComfyUI and the models it downloads.
- An agent already set up on your RTX Spark. If you don’t have one yet, follow the introductory agent playbook first, then come back here.
- The creative app you want your agent to work with, installed on your RTX Spark. What the agent can do depends on the app.
