-
-
dashboard
Inspiration
Many students and builders struggle with urgent deadlines because they know what they want to finish, but they do not know how to break it into clear action steps. During hackathons, this becomes even harder because participants need to manage coding, GitHub, deployment, demo videos, and final submission all within a short time.
DeadlinePilot AI Agent was built to solve this problem by turning any urgent goal into a structured, trackable action plan.
What it does
DeadlinePilot AI Agent helps users enter a high-priority goal and a deadline. The agent then uses Gemini to generate a 6-step task plan with priorities, descriptions, deadlines, and status.
The generated tasks are stored in MongoDB, where the app keeps track of goals, task progress, and agent action history. Users can view their task dashboard, mark tasks as completed, track progress, and generate Devpost-ready submission content.
How we built it
The project was built using Streamlit for the user interface, Gemini API for AI reasoning and task generation, and MongoDB Atlas for storing goals, tasks, progress, and logs.
Gemini acts as the planning brain of the agent. MongoDB acts as the memory and persistence layer. Together, they allow the project to go beyond a simple chatbot by actually creating, storing, retrieving, and updating real workflow data.
MongoDB Integration
MongoDB is used to store:
- User goals
- Generated task plans
- Task completion status
- Agent action logs
- Previous goal history
This makes the agent stateful and useful across multiple interactions. Instead of only giving text responses, the agent saves structured data and allows users to continue tracking their progress.
Challenges we ran into
One challenge was making the agent reliable while keeping the project simple enough for a hackathon MVP. Another challenge was handling API keys, MongoDB connection issues, and task storage safely without exposing secrets.
We also focused on making the dashboard clean and professional so that the user experience feels useful, not just experimental.
Accomplishments that we're proud of
We created a working AI agent that can understand a deadline-based goal, generate a structured plan, store it in MongoDB, update progress, and generate Devpost submission content.
The project successfully demonstrates how an AI agent can help users move from confusion to action.
What we learned
We learned how AI agents become more powerful when connected to real tools like databases. Gemini provides reasoning, while MongoDB gives the agent memory and persistence.
This helped us understand how agents can be designed not just to answer questions, but to perform useful actions.
What's next
Future improvements include calendar reminders, email notifications, team collaboration, analytics, voice input, and deeper integration with Google Cloud Agent Builder and MongoDB MCP.
Log in or sign up for Devpost to join the conversation.