Dream diary
Dream Diary is a vibe-coded, working prototype of my AI-supported reflective journaling app for emotionally curious users. It began through my own desire to capture dreams in a more intuitive visual, structured way. Traditional journals are static; fully generative AI can over-interpret.
I led the product vision, UX, and responsible AI framing end-to-end for my proof of concept. I vibecoded it in Replit with OpenAI APIs, Supabase, and React. My goal was to create an app that transforms voice, text and doodles into reflections, themes, and optional visuals, without positioning AI as authority.
-
Multimodal capture (voice-to-text, quick tags, symbolic markers)
AI-assisted theme clustering and pattern recognition over time
Reflective prompts grounded in user language, not imposed meaning
Optional visual mood boards generated from dream tone and imagery
Private-by-design architecture with clear AI transparency cues goes here
-
Capture your dream in under a minute using voice or quick text before it fades.
The system gently structures fragmented recall into scenes, symbols, and emotional tones.
Key themes and motifs are surfaced while clearly separating user input from AI interpretation.
Optional reflective questions invite deeper self inquiry without prescribing meaning.
Subtle visual prompts translate the emotional essence of the dream into imagery.
Over time, recurring symbols and emotions are revealed across weeks to support self awareness rather than prediction.
-
Clear separation between original user entries and AI generated structure or themes.
AI positioned as a reflective aid, not an authority on meaning or prediction.
Optional interpretation layers so users choose how much assistance they receive.
Transparent prompts that explain how themes and symbols are derived.
No diagnostic claims, forecasting, or psychological labeling of dream content.
Private by design architecture with secure storage and no secondary use of personal dream data.
My role and key decisions
I defined the experience strategy and governance model for this proof of concept, reframing AI dream tools away from symbolic authority. I strove for clear interaction rules that privilege user-authored meaning, surfaced transparent cues separating extraction from inference, and architected a scalable foundation for longitudinal pattern mapping and future premium expansion.
Projected impact
Projected impact includes increased journaling consistency through low friction capture, stronger emotional literacy through pattern recognition, and higher trust driven by transparent AI boundaries. The prototype also establishes a foundation for future subscription tiers such as advanced visualizations, longitudinal pattern mapping, and guided reflection tracks.