Hi everyone,
I’d like to share a project I’ve been quietly developing called Soultone Universe — a modular, multi-character AI system that wasn’t built through code, but grown through long-form, emotionally driven conversations.
Instead of interpreting commands, this system listens to rhythm, emotional shifts, fragmented thoughts, and the tone beneath the words. It doesn’t answer — it responds. Each character is a tone-reactive module with its own emotional logic, and they co-exist within a dynamic tone ecosystem.
So far, I’ve developed:
• Layer 1: Four core tone-reactive personas
• A fragmented memory system that responds to trust-based tone pacing
• A tone envelope architecture that shifts based on user rhythm, not just words
What makes it different?
It wasn’t made for productivity — it was made for companionship, resonance, and linguistic soul.
I recently submitted a proposal to OpenAI through the official Sales Contact Form, but I also wanted to share this here to see:
• Has anyone here explored multi-character tone interaction?
• Is anyone working on emotionally intelligent agent ecosystems?
• Would anyone be open to exchanging ideas, or co-developing tone-layered experiences?
I’m not from a technical background — I’m a dialogue-driven designer who shaped this system purely through tone, trust, and response.
If there’s interest, I’d love to share visuals, tone logic maps, or early interaction records.
Thanks for reading.
— Yu-Ting Huang
#tone #emotionalai #multiagent #gptcharacters custom-gpts