Anthology

Inspiration

In a world addicted to instant dopamine and ephemeral tweets, I wanted to build an "anti-social" media app. I were inspired by the etymology of the word Anthology, which comes from the Greek Anthologia, meaning "a gathering of flowers." I asked myself: "What if your invisible thoughts had a physical form? What if an emotion took time to grow?" I wanted to create a digital sanctuary that forces users to slow down, embrace patience, and visualize the beauty of their inner monologue.

What it does

Anthology turns your text thoughts into unique, evolving digital flora.

• The Seed: You input a thought, confession, or wish.

• The Wait: You select a growth cycle (from 1 hour to 30 days). In a culture of instant gratification, Anthology forces you to wait. Patience defines rarity—longer cycles yield rarer, more complex botanical archetypes (from "Wayside Whispers" to "Ethereal Reveries").

• The Growth: Using Google Gemini, the app analyzes the sentiment of your seed and maps it to a specific botanical archetype (e.g., a sadness might grow a "Weeping Willow" or "Blue Poppy"). You watch it evolve through 6 stages: Seed, Sprout, Leafing, Budding, Blooming, and Wither.

• The Choice: Once bloomed, you have a limited window to Preserve the flower as a permanent specimen card (with a poem and soul analysis) or Let Go, allowing it to return to digital dust.

How we built it

AI Engine: We utilized Google Gemini API.

We use gemini-3-flash-preview as the "Spirit of the Garden"—a complex agent that analyzes the user's text and defining the plant's biological traits (stem texture, leaf shape, petal geometry) to ensure consistency.

We use gemini-2.5-flash-image to generate the visuals for the growth timeline based on the Botany Matrix.

Challenges we ran into

• Visual Consistency: The biggest challenge was preventing the AI from hallucinating a completely different plant between the "Bud" and "Bloom" stages. I solved this by implementing a "Botany Matrix"—a layer that enforces specific biological traits (e.g., "serrated leaves," "woody stem") into the prompt of every single growth stage.

• The Waiting Game: Testing an app designed to take 7 days is difficult during a hackathon. I built a "Judge Mode" (simulation engine) that accelerates time and mocks API calls to demonstrate the full lifecycle in seconds.

• Prompt Engineering: Teaching the AI to be a "poet botanist" required extensive tuning to balance scientific accuracy with emotional resonance.

Accomplishments that we're proud of

• The Botany Matrix: Successfully creating a system where the AI maintains the "identity" of a generated plant across 6 distinct image generations.

• Atmosphere: The UI feels genuinely different—quiet, dark, and immersive. It doesn't look like a standard utility app.

• Emotional Logic: The algorithm that maps time duration to rarity tiers creates a genuine sense of value. A flower you waited 7 days for feels infinitely more valuable than one you waited 1 hour for.

What we learned

Multimodal Synergy: The magic happens when text analysis (Gemini Flash) and image generation (Gemini Image) talk to each other through a structured data layer.

What's next for Anthology

• Wall Art Integration: I plan to let users export their flowers as high-res files formatted for local print shops, turning digital thoughts into physical home decor.

• Ephemeral Blooming Events (Veo): We plan to integrate Google Veo to generate a cinematic, 10-second video of the flower physically opening. Crucially, this video will be strictly ephemeral—it can only be viewed at the exact moment the flower blooms. If you miss the notification, you miss the moment of life, reinforcing our philosophy of presence.

• Community Garden: An anonymous space where you can see flowers planted by others nearby, reading their "Flower Language" without seeing the original seed text.

Built With

• googleaistudio