TaskPilot AI

• 

  Home Page

Inspiration

Students and professionals often receive tasks through scattered sources like screenshots, emails, PDFs, and rough notes. This unstructured information makes planning overwhelming and leads to missed deadlines. I wanted to build an AI system that doesn’t just respond to questions, but actually thinks like a planning assistant and turns chaos into clarity.

What it does

TaskPilot AI is an intelligent productivity agent that converts messy inputs into structured action plans.

Users can upload screenshots, PDFs, or text notes, and the AI:

1. Extracts tasks
2. Breaks complex work into smaller steps
3. Prioritizes based on urgency and effort
4. Generates a clear, actionable to-do plan

It acts like a smart work planner, not just a chatbot.

How I built it

I built a web app using React.js for the frontend and Python (Flask) for the backend. The Google Gemini API powers the core intelligence.

Gemini processes multimodal inputs (text + images/PDFs), performs task extraction and reasoning, and returns structured task plans in JSON format. The backend formats this data and displays it in an organized task dashboard for users.

Challenges I ran into

One major challenge was converting unstructured content into reliable structured output. Ensuring Gemini consistently returned properly formatted task lists required careful prompt design and output structuring. Another challenge was balancing advanced AI reasoning with a simple and intuitive user interface within a short hackathon time.

Accomplishments that I'm proud of

I built a working AI agent — not just a chatbot — that demonstrates real reasoning and planning ability. TaskPilot AI successfully transforms raw information into meaningful action plans, showing how generative AI can function as a productivity assistant rather than only a conversational tool.

What I learned

I learned how powerful Gemini can be when used as a reasoning engine instead of just a text generator. Prompt engineering, structured outputs, and defining clear AI roles were key to building a reliable agent. I also gained experience in designing AI-first applications within tight time constraints.

What's next for TaskPilot AI

Next, I plan to add calendar integration, deadline tracking, and adaptive scheduling based on user workload. I also want to introduce reminders, progress tracking, and team collaboration features to turn TaskPilot AI into a complete AI-powered productivity platform.

Built With

• css3
• google-cloud
• googlegeminiapi
• html5
• json
• python
• react.js