FlowMind AI

Here's your complete Devpost "About the project" section — ready to copy-paste:

Inspiration

Most people fail at their goals not because they lack motivation, but because they lack a clear plan. We've all stared at a big goal — "launch my startup," "build my portfolio," "ace my exam" — and felt paralyzed, not knowing where to begin. Existing productivity tools just give you a blank box and say "add your tasks." That's not help. We wanted to build something that actually thinks with you — so we built FlowMind AI.

What it does

FlowMind AI is an AI-powered productivity co-pilot. You type any goal in plain English, and it instantly breaks it down into a prioritized, time-estimated task list. It then builds a day-by-day schedule automatically. An AI coach — fully aware of your goal and current progress — is available anytime to help you overcome blockers, reprioritize, and stay on track. It also tracks your progress in real time with visual stats and a completion bar.

How we built it

We built FlowMind AI as a full-stack React web application powered by the Claude API. The app has three core views: Task Planner, Daily Schedule, and AI Coach. For task generation, we engineered a structured prompt that reliably returns a JSON array of tasks with priority, time estimate, and day assignment. For the AI coach, we use a multi-turn conversational prompt where the user's full task list and goal are dynamically injected as context on every call — making advice specific rather than generic. The app is deployed on Vercel.

Challenges we ran into

Getting the Claude API to return consistent, parseable JSON without markdown fences or extra explanation required significant prompt engineering — we iterated through many versions before landing on a reliable format. Managing the AI coach's context window efficiently across a multi-turn conversation was another challenge, especially keeping responses fast and relevant. We also had to carefully design the UX around AI latency — adding typing indicators and smooth loading states so the experience felt natural rather than broken.

Accomplishments that we're proud of

We built a genuinely working, polished product in 48 hours — not a mockup, not a concept, but a live tool you can use right now. The combination of structured AI task generation and a context-aware conversational coach in one seamless interface is something that doesn't exist in any current productivity app. We're also proud of the prompt engineering behind task generation — getting reliable structured output from a language model consistently is harder than it sounds.

What we learned

We learned that prompt engineering is a real engineering discipline — small wording changes produce dramatically different outputs. We also learned how powerful context injection is for making AI assistants feel genuinely useful rather than generic. On the product side, designing UX around AI latency taught us a lot about perceived performance and how loading states can make or break an experience.

What's next for FlowMind AI

We plan to add Google Calendar integration so tasks are scheduled directly into your real calendar. Team collaboration is on the roadmap — share a goal, assign tasks to teammates, and get AI coaching as a group. We're also building a mobile app, AI-powered time blocking based on your existing calendar availability, and daily email or Slack digests of your plan. Long term, FlowMind AI could evolve into a full AI chief-of-staff for individuals and small teams.

Built With

• ai
• claud