Gritto

• 

  Daily task list and goal list

• 

  Preview the generated goal

• 

  On boarding

• 

  Chat with Gritto to plan your goal

• 

  Architecture Diagram

Inspiration

Many people set goals but struggle to stay consistent. We wanted to create an AI companion that not only plans but adapts with you — turning ambition into achievable, balanced progress.

What it does

Gritto helps users chat with an AI to create, refine, and track goals, automatically scheduling tasks into their available time. It also lets users sort priorities, check off to-dos, and reflect on progress.

How we built it

We used Kotlin Multiplatform for the mobile client, Express + Firestore for the backend, and Google’s ADK with Gemini for our goal-planning agent — all deployed on Cloud Run for a fully serverless architecture.

Challenges we ran into

• Synchronizing session states between backend and agent.
• Designing structured LLM outputs that stay JSON-valid.
• Balancing user availability with AI-generated schedules.

Accomplishments that we're proud of

• End-to-end AI goal creation flow from chat to schedule.
• Multi-agent reasoning deployed serverlessly on Cloud Run.

What we learned

• How to orchestrate multi-agent reasoning with Google ADK.
• Effective schema design for dynamic AI planning.
• The power of serverless scalability for rapid iteration.

What's next for Gritto

We plan to add voice reflection, emotional tone analysis, and a web dashboard for insights — evolving Gritto into a truly adaptive personal growth ecosystem.

Built With

• adk
• cloudrun
• compose
• gcp
• kotlin