Inspiration

In an era dominated by AI efficiency, a shadow of existential anxiety looms over us. People fear being replaced by machines, and daily life often feels reduced to a series of repetitive, meaningless tasks. We wanted to flip the script: If AI is the one "taking" our jobs, can it also be the one "giving" us back our inner peace? Our inspiration comes from the concept of "Art Therapy." We believe that every fleeting emotion—no matter how trivial—deserves to be seen. By transforming the "small certainties" of daily life into unique visual memories, we help users reconnect with their inner selves in a world that feels increasingly fragmented.

What It Does

Petal is an AI-powered art therapy companion. Users record their daily moods or thoughts through simple text or voice. Our AI then translates these abstract emotions into a personalized piece of digital art. It’s not just an image; it’s a visual diary—a private gallery where your feelings are curated into a collection of "Emotional Artworks," turning ephemeral moments into permanent, aesthetic memories.

How We Built It

We utilized Large Language Models (LLMs) as the "emotional translator" and combined them with state-of-the-art Image Generation models. The core of our build was creating a bridge between sentiment analysis and artistic prompt engineering. We spent significant time building a pipeline that interprets the nuance of human emotion and maps it to specific artistic movements, color palettes, and textures.

Challenges We Ran Into

Our biggest challenge wasn't the code, but the project target and scope. We initially spent hours brainstorming a fitness and diet tracker for pregnant women. However, we couldn't relate to the experience, and the passion just wasn't there. We had to navigate the struggle between "what we think the market wants" and "what we are actually passionate about." This conflict was exhausting but led us to change our direction.

  1. Challenge 1: Language Excludes the People Who Need Expression the Most When we set out to make emotional reflection accessible to everyone, we immediately confronted a core contradiction: the people who most need to process their emotions are often the least able to articulate them. Emotional distress is frequently pre-verbal — it lives in heaviness, in color, in a vague sense that something is wrong, long before it becomes a sentence. Existing journaling tools ask users to write, which places the highest cognitive burden on people at their most depleted moments. We addressed this by making art — not text — the primary medium of expression. Visual and creative expression has decades of clinical research behind it precisely because it bypasses the effort of language. In Canvas, users don't need to find the words. They share a photo, a voice note, or a few fragments of their day, and receive back a painting and a poem that reflects their emotional world. The image holds what language couldn't. This shift in medium is not aesthetic — it is the core of our social impact thesis: that emotional reflection should be accessible to anyone, regardless of their ability to articulate what they feel.
  2. Challenge 2: Emotional Habits Are Easy to Start and Impossible to Keep The second challenge we faced is one that undermines almost every mental health tool ever built: sustainability. People do not abandon emotional wellness practices because they stop caring — they abandon them because the practice feels like effort without visible return. A blank journal page offers no feedback. Progress in emotional health is invisible. And invisible progress feels like no progress at all. We addressed this through the design of the ritual itself. In Canvas, every day ends not with a completed form, but with a piece of art that is uniquely yours — generated from your life, your feelings, your aesthetic choices. The calendar fills slowly with color and imagery. The monthly gallery becomes a visual autobiography you can return to. The AI-generated monthly reflection gives users a narrative of their own emotional journey over time — transforming what is usually an invisible, private struggle into something they can see, hold, and feel proud of. When emotional growth becomes visible, the habit becomes worth keeping.

Accomplishments That We’re Proud Of

We are incredibly proud of the Aesthetic Evolution of our AI. Early on, the generated images were cold, generic, and frankly, didn't match the user's mood at all. We refused to settle for "AI-looking" art. Through iterative refinement—training the model to mimic specific artistic styles and learning from our own curated aesthetic tastes—we successfully moved the needle. Now, the images aren't just "generated"; they are "composed" with a sense of depth and historical resonance that we truly love.

What We’ve Learned

This hackathon was a lesson in Product Taste. We realized that building something new isn't just about technical skill; it’s about "soaking" in the product ecosystem long enough to catch the right trends. We learned that:

  1. Authenticity Trumps Utility: If you don't relate to your product, no one else will.
  2. The "Pivot" is a Strength: Spending time on "detours" isn't a waste if it leads to a project you actually care about.
  3. AI is a Mirror: The hardest part of building an AI product is figuring out exactly who you are building it for and what "value" really means beyond just efficiency.

What’s Next for Our Product

As we look forward, our roadmap is focused on Safe Integration and Professional Support:

  1. B2B for Therapists: We are exploring a "Professional Suite" where Canvas helps therapists organize patient archives and emotional trends, providing them with a visual "mood map" to better understand their clients' progress between sessions.
  2. Mitigating "Negative Feedback Loops": We are researching ways to ensure that users in a negative headspace aren't met with "triggering" imagery. Our goal is to refine the AI to create Cathartic Art—art that acknowledges sadness but renders it with a sense of beauty and hope, rather than reinforcing the gloom.
  3. Safety-First Design: To avoid legal risks and ensure user safety, Canvas will remain a tool for self-awareness and reflection rather than prescriptive medical advice, maintaining a clear boundary between "healing art" and "clinical therapy."

Built With

  • chatgpt
  • claude
  • lovable
Share this project:

Updates

Submission history