PromiseLoop

main dashboard
ai that helps to undr stand
promise list
mcp

Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for PromiseLoop

Inspiration

Promises are easy to make but easy to forget. We built PromiseLoop to help people keep track of their commitments without manually filling out forms or maintaining complicated to-do lists.

What it does

PromiseLoop uses Gemini AI to understand promises expressed in natural language, extract tasks and deadlines, and prepare commitment drafts. Users confirm before anything is saved. A self-hosted MCP server exposes tools to manage commitments, dependencies, completion, and next actions.

How we built it

We developed a React and Vite frontend, a FastAPI backend, and a persistent SQLAlchemy database. Gemini powers natural-language understanding, while the Model Context Protocol (MCP) connects the application to six commitment-management tools over Streamable HTTP.

Challenges we faced

One major challenge was making natural-language promises actionable without inventing missing details. We also focused on preventing unconfirmed commitments from being saved and verifying real MCP communication rather than relying solely on simulated tool calls.

What we learned

We learned how to integrate AI with structured workflows, implement MCP over Streamable HTTP, preserve user control through confirmation gates, and test the complete journey from conversation to persistent commitment.