Hackathon Helper is an AI-powered orchestration tool designed to eliminate the "cold-start" problem of hackathons. By automating the transition from a dense Devpost overview to a structured technical blueprint, it ensures developers spend their first hour coding instead of configuring.
Success in a hackathon is about consistency and reliability. This tool is built to help you meet every requirement and ship a functional, polished project by identifying a clear, achievable path through the requirements from the very start.
- Strategic Analysis: Extracts mandatory requirements and proprietary platform constraints (like DigitalOcean or GitLab integrations).
- Dynamic Scaffolding: Automatically generates
.gitignoreandrequirements.txtbased on the AI's prediction of your project's specific library needs. - Creative Brainstorming: Suggests achievable project concepts tailored to the extracted hackathon criteria.
- Professional Dashboard: A Flask-based web interface (
app.py) that renders your blueprint into a polished, browser-based report for high-quality presentation.
The project is organized to support both the multi-agent logic and the professional presentation layer:
DIGITALOCEAN_GRADIENT_AI_HACKATHON
├── app.py # Flask Web Interface (Professional Reporting)
├── main.py # Core Multi-Agent Logic (LangGraph + Gradient)
├── devpost_input.txt # The "Source of Truth" (Paste Hackathon text here)
├── hackathon_report.md # Markdown Output
├── hackathon_report.json # Data Output
├── output/ # Generated Project Sandboxes
└── docs/ # Documentation & Logs
The system uses a LangGraph multi-agent workflow powered by Gradient AI:
-
Analyst Node: Maps out the mandatory "Must-Builds" and technical constraints.
-
Scaffolder Node: Reasons about Python dependencies and creates the local environment files.
-
Creative Node: Identifies valid, high-potential ideas that fit the specific hackathon scope.
-
Flask Layer (
app.py): Dynamically executes the workflow and renders the results in a clean, professional dashboard.
Ensure your .env file contains your DIGITALOCEAN_API_TOKEN and GRADIENT_MODEL_ACCESS_KEY.
pip install -r requirements.txtPaste the text from the hackathon overview page on DEVPOST into devpost_input.txt.
Launch the web dashboard. This single command triggers the multi-agent workflow, generates your project scaffolding, and prepares your report:
python app.py- Browser Report: Navigate to http://localhost:8081 to view the professionally rendered analysis and project concepts.
- Physical Scaffolding: Check the
/output/directory. The agent will have automatically created a new folder named after the hackathon containing your customized.gitignoreandrequirements.txt. - Data Exports: Find the raw
hackathon_report.jsonandhackathon_report.mdin the root directory for your records.
Building this tool demonstrated the power of Multi-Agent Systems (MAS). By breaking down the problem into specialized roles (Analyst, Scaffolder, Creative), we achieved much higher accuracy and utility than a single LLM prompt could provide.
GitLab Integration: Auto-committing scaffolded files to a new repository via webhooks.
Cloud Deployment: Migrating the local dashboard to a persistent cloud environment.
Multi-Language Support: Expanding scaffolding beyond Python to include Node.js, Rust, and Go.
Built for the DigitalOcean & Gradient AI Hackathon.