SpeakEase

Architecture Diagram

Inspiration

Inspired by our experiences as introverts and our personal connections with individuals on the autism spectrum, we created SpeakEase to empower users with confident communication and emotional resilience.

What it does

SpeakEase provides: Calm Space – immediate emotional support during stressful moments. Paraphrasing Tools – one simplifies complex language for better understanding, the other enhances clarity so users can express themselves more effectively.

How we built it

We use agent development kit in vertex AI to develop the multi agent framework, which includes mcp server, and is deployed on agent engine. We use flutter to develop our frontend, and fastapi to develop the backend. Firebase on GCP is developed for authentication, and google cloud firestore is developed to store our data. MCP toolbox is used for getting data from PostgreSQL database. All of these are integrated for real-time, personalized support.

Challenges we ran into

Deployment and infrastructure setup were key hurdles under tight deadlines.

Accomplishments that we're proud of

✅ Successfully developed and deployed a multi-agent framework using Vertex AI's Agent Development Kit and Agent Engine for scalable and adaptive communication support. ✅ Created a seamless user experience with a cross-platform Flutter frontend and a high-performance FastAPI backend. ✅ Integrated real-time emotional support features such as Calm Space to assist users during moments of stress or overwhelm. ✅ Implemented dual paraphrasing tools for both simplifying complex input and enhancing self-expression clarity. ✅ Enabled secure and user-friendly authentication using Firebase Authentication on Google Cloud Platform. ✅ Leveraged Google Cloud Firestore for scalable, real-time storage of user data and preferences. ✅ Built a robust backend data pipeline using MCP Toolbox for efficient access to a PostgreSQL database.