Update — Migration to karaoke.vision & Gemini Inference
We've shipped a major update based on community feedback. Two key changes:
1. Rebranded under karaoke.vision
SETMIND now operates under the karaoke.vision umbrella. The interface, routing, and product surface have been updated to reflect the new identity.
2. Switched inference from Nemotron to Gemini
Several users reported high latency on AI features — set planning, crowd analysis, and AI Mixer transitions were taking too long to feel usable in a live booth. We heard you.
We've migrated the entire inference layer from the Crusoe-hosted Nemotron model to Google Gemini. The result is significantly faster response times across all AI capabilities, with thinking disabled to dedicate the full token budget to the answer. A model fallback ladder with retry logic keeps things stable under load.
What this means for you:
- AI features now respond noticeably faster
- Connect your own Google Gemini API key in the in-app settings panel (stored locally in your browser)
- A status indicator confirms the connection once your key is active
- Set Planner, Crowd Pulse, AI Mixer, and Auto-Queue are all fully operational on the new backend
Thanks to everyone who flagged the latency — this one's for you.
Log in or sign up for Devpost to join the conversation.