MakeMeBetter AI

MakeMeBetter AI is a complete, deployed productivity application — goals, habits, and tasks in one system — with an AI integration that goes well past a chat box: an MCP server that lets assistants like Claude actually operate your board.

The problem

Most trackers treat tasks as a flat list. They don’t connect a task to the habit that builds it or the goal it serves, so the thing you’re doing today gets disconnected from the thing you’re trying to become. And none of them let an AI assistant do the work with you — read your board, add a subtask, mark progress.

What I built — end to end

• A real hierarchy — goals contain habits, habits and tasks nest into subtasks, all with progress tracking and completion state.
• Plan tiers — free and pro plans with per-plan quotas, i.e. a product shaped for real users, not a demo.
• An MCP server — the standout piece. AI assistants can list, create, update, and complete goals/habits/tasks directly through the Model Context Protocol, with the user’s own board as the source of truth. Productivity software an agent can genuinely drive.
• Deployed and live — running in production, backed by a persistent database and authenticated per-user data.

Where PolicyChat AI was a prototype to prove an idea, MakeMeBetter AI is the full build — the piece I take furthest, from data model to deployment to an AI-native interface.