W5 · TETRA SAIAR
Interactive Generative AI
📝 Description
In this workshop, participants explore how to use generative AI in real time, with a focus on fast experimentation and reusable workflows. TouchDesigner is the main environment: we cover the basics of TOPs, CHOPs and DATs, working with parameters, and preparing a network for performance mode. From there, we build real-time chains that connect live inputs (camera, audio, controllers, sensors) to responsive visuals on screen. Participants learn how to plug in AI models for live content generation and transformation, and how to combine these elements into practical setups for shows and installations.
🛠️ Tools & Tech
- AI models: StreamDiffusion (and comparable real-time diffusion models)
- Software: TouchDesigner
- Optional tools: ComfyUI / AI toolkits for connecting external models
- Hardware: Laptop
🌍 Applied Scenario
Live shows, interactive installations and brand activations: AI-driven visuals that move with sound, react to audience movement or are controlled via sliders, MIDI, or sensors in real time.
🎯 Outcomes
- Learn TouchDesigner basics: working with TOP / CHOP / DAT operators, parameters, exporting and performance mode.
- Understand the end-to-end chain from real-time input (camera / audio / controller / sensors) to visual response.
- See how to slot generative AI models into a real-time visual pipeline.
🗂️ Workshop Flow
- Introduction (30 min) — Context and inspiration
- TouchDesigner foundations (60 min) — TOP / CHOP / DAT basics, experimenting with workflows
- Realtime input → visual chain (1u)
- AI in the loop (60 min) — AI models for live content
- Hands-on exercise (1,5 hours)
- Show & discuss (30 min) — Share outcomes & gather feedback
Interested in a workshop?
These workshops can be tailored to your team and booked on request. Get in touch with Giliam to discuss dates, scope and what would fit your organisation.
Contact Giliam