The problem
Most AI assistants are optimized for answers, productivity, and efficiency. But when people vent, they do not want instant solutions, robotic empathy, or over-optimized responses. They want space, understanding, and natural flow.
Project
Venting with Omnia
Venting with Omnia is a voice-first AI companion designed for emotional expression, a space where users can talk freely, process thoughts, and feel heard without judgment.
The approach
Venting with Omnia is designed around emotional UX, not just intelligence. The system prioritizes listening over solving, pacing over speed, and tone over information so the conversation can unfold naturally.
What makes it technically interesting
- Voice-first architecture reduces friction compared to typing and encourages more natural expression.
- Conversational pacing is intentionally balanced to avoid overloading the user.
- Tone shaping guides responses to feel calm, grounded, and non-judgmental without exaggerated empathy.
- Minimal UI philosophy keeps the interface out of the way so the conversation stays central.
- State-aware responses track emotional context across turns and avoid abrupt shifts in tone or direction.
Who it’s for
- People who want to process thoughts out loud.
- Users looking for a non-judgmental space.
- Anyone who prefers speaking over typing.
What this demonstrates
This project explores how AI can function as a companion, not just a tool. It shows that good AI interaction is not just about intelligence. It is about timing, tone, and restraint.
What I learned
Helping someone think is different from answering their question. In emotional contexts, the instinct to solve too quickly is often wrong. Designing for this meant resisting over-response, allowing space, and prioritizing how something is said over what is said.
Why this matters
As AI becomes more integrated into daily life, emotional interaction will matter more. Systems that can listen well, respond appropriately, and respect human pacing will feel significantly more natural and trusted.