Type
Independent
Role
Role
UI / UX Design
UI / UX Design
Ideation
Ideation
Vibe Coding
Vibe Coding
Prototyping
Timeline
Timeline
June - Aug 2025 (8 weeks)
June - Aug 2025 (8 weeks)
Tools
Tools
Figma
Figma
The Space
The Space
An AI auto-tagged chat interface where conversations become space, not a disappearing feed.
Overview
Overview
Modern chat interfaces are built on a disappearing feed — messages slide upward, context dissolves, and meaningful conversations flatten into history. Only the last few messages remain visible. Everything else requires scrolling, searching, or guessing.
Modern chat interfaces are built on a disappearing feed — messages slide upward, context dissolves, and meaningful conversations flatten into history. Only the last few messages remain visible. Everything else requires scrolling, searching, or guessing.
For this project, I explored a different question:
What if conversations didn’t scroll away—
but accumulated into a space you could return to?
What if conversations didn’t scroll away—
but accumulated into a space you could return to?
What if conversations didn’t scroll away—
but accumulated into a space you could return to?
I set out to reimagine chat as a persistent, spatial memory system—one where AI doesn’t replace conversation, but helps organize it.
Early Experiments
Early Experiments
I explored with AI prototyping tools like Figma Make and Replit, using gamification and a literal “conversation room” to make chats feel active and tangible. But the ideas were too heavy and distracting. These tests helped me realize the system needed to be spatial but not literal.
Development
Development
Later, I experimented with ways to visualize conversation context more abstractly. I explored how variations in color, scale, and typographic weight could represent emotional tones, intensity, and shifts in mood over time. These prototypes tested different mappings—bars, grids, blocks, and expressive text layouts—to understand how emotion could be surfaced without relying on literal rooms or gamified elements. This phase helped me identify the visual language and semantic structures that would later support a more fluid, spatial interpretation of conversation memory.
Later, I experimented with ways to visualize conversation context more abstractly. I explored how variations in color, scale, and typographic weight could represent emotional tones, intensity, and shifts in mood over time. These prototypes tested different mappings—bars, grids, blocks, and expressive text layouts—to understand how emotion could be surfaced without relying on literal rooms or gamified elements. This phase helped me identify the visual language and semantic structures that would later support a more fluid, spatial interpretation of conversation memory.
Vibe Coding
Vibe Coding
I prompted a demo in Figma Make, demonstrating the feature of pulling key topics into the side bar.
I prompted a demo in Figma Make, demonstrating the feature of pulling key topics into the side bar.
Design System
Design System
Typography
Typography
Noto Sans provides a clear, flexible typographic voice that can scale from subtle metadata to expressive, oversized topic words.
Noto Sans provides a clear, flexible typographic voice that can scale from subtle metadata to expressive, oversized topic words.
Primary Colors
Primary Colors
A violet-based palette creates a calm, atmospheric canvas that supports depth, contrast, and the project’s memory-focused tone.
A violet-based palette creates a calm, atmospheric canvas that supports depth, contrast, and the project’s memory-focused tone.
Secondary Colors
Secondary Colors
Bright accent colors give each topic a distinct visual identity, making emotional and thematic patterns instantly recognizable across the interface.
Bright accent colors give each topic a distinct visual identity, making emotional and thematic patterns instantly recognizable across the interface.
Icons
Icons
Simple, geometric icons pair with each topic as semantic anchors, scaling cleanly from small tags to dense visual clusters.
Simple, geometric icons pair with each topic as semantic anchors, scaling cleanly from small tags to dense visual clusters.
Final Outcome
Final Outcome
Flow #1
Flow #1
Start Conversation with a Friend
Start Conversation with a Friend
Every friend becomes a visual profile built from your shared conversation history—
where color bars, streaks, and topic patterns reveal the emotional tone of the relationship at a glance.
Every friend becomes a visual profile built from your shared conversation history—
where color bars, streaks, and topic patterns reveal the emotional tone of the relationship at a glance.
Flow #2
Flow #2
New Conversation Topic Added
New Conversation Topic Added
The moment of sending a message and watching the system update in real time—
AI analyzes tone, assigns topic color, and integrates the message into the spatial display without relying on a scrolling feed.
The moment of sending a message and watching the system update in real time—
AI analyzes tone, assigns topic color, and integrates the message into the spatial display without relying on a scrolling feed.
Flow #3
Flow #3
Archived Conservation Topic Tags
Archived Conservation Topic Tags
This reveals how AI-generated tags surface alongside messages, forming topic clusters that grow, shift,
and reorganize as the conversation evolves, turning the dialogue into a navigable semantic map.
This reveals how AI-generated tags surface alongside messages, forming topic clusters that grow, shift,
and reorganize as the conversation evolves, turning the dialogue into a navigable semantic map.
Flow #4
Flow #4
Add New Friend to Current Chat
Add New Friend to Current Chat
New connections enters the system, immediately generates a lightweight profile that begins
tracking topics, emotional tones, and conversational patterns as soon as the first messages are exchanged.
New connections enters the system, immediately generates a lightweight profile that begins
tracking topics, emotional tones, and conversational patterns as soon as the first messages are exchanged.
Reflection
Reflection
Working with AI fundamentally reshaped my design process. It expanded the range of prototypes I could explore in a short amount of time and allowed me to test unconventional metaphors—rooms, clusters, maps—long before committing to a direction. AI shifted my role from drawing individual screens to designing an underlying system of meaning, where data, context, and emotional tone all play structural roles.
This project also made me rethink what chat is: not just a stream of messages, but a form of shared memory. It pushed me to see digital presence as something spatial rather than linear, and conversations as relationships that accumulate texture over time. I came away understanding that interfaces can hold emotional history—not just display text—and that AI can help reveal the structure beneath our relationships without overwhelming the human experience.
Working with AI fundamentally reshaped my design process. It expanded the range of prototypes I could explore in a short amount of time and allowed me to test unconventional metaphors—rooms, clusters, maps—long before committing to a direction. AI shifted my role from drawing individual screens to designing an underlying system of meaning, where data, context, and emotional tone all play structural roles.
This project also made me rethink what chat is: not just a stream of messages, but a form of shared memory. It pushed me to see digital presence as something spatial rather than linear, and conversations as relationships that accumulate texture over time. I came away understanding that interfaces can hold emotional history—not just display text—and that AI can help reveal the structure beneath our relationships without overwhelming the human experience.
Next Step
Next Step
I plan to explore how this system could adapt to group conversations, long-term relationship timelines, and richer emotional modeling. I’m also interested in building an interactive prototype that tests how people navigate conversational memory in real time, and how AI might surface patterns or reminders that support healthier, more intentional communication.
I plan to explore how this system could adapt to group conversations, long-term relationship timelines, and richer emotional modeling. I’m also interested in building an interactive prototype that tests how people navigate conversational memory in real time, and how AI might surface patterns or reminders that support healthier, more intentional communication.