HackForge AI is a futuristic Multi-Agent AI platform that helps developers, students, and hackathon teams generate complete hackathon-ready project plans automatically using local AI models powered by Ollama and Phi3.
Built during the Build with AI – Agent Builder Camp organized by GeeksforGeeks × Google for Developers.
✅ Multi-Agent AI Workflow
✅ AI-Powered Project Idea Generation
✅ Tech Stack Recommendation System
✅ GitHub README Generator
✅ Modern Futuristic UI
✅ Local AI Inference using Ollama + Phi3
✅ Streamlit Interactive Dashboard
✅ Export Generated Outputs
✅ Lightweight & Beginner Friendly
✅ Offline AI Support
Generates:
- Hackathon project ideas
- Startup concepts
- Problem statements
- Target users
- Core features
Suggests:
- Frontend technologies
- Backend frameworks
- APIs
- AI tools
- Deployment stack
Automatically generates:
- GitHub README structure
- Installation instructions
- Project documentation
- Setup guide
| Technology | Purpose |
|---|---|
| Python | Core Development |
| Streamlit | Frontend Dashboard |
| Ollama | Local LLM Runtime |
| Phi3 | Lightweight AI Model |
| Requests | API Communication |
git clone https://github.com/shreya975/hackforge-ai.git
cd hackforge-aipip install -r requirements.txtDownload Ollama:
ollama run phi3Wait until you see:
>>> Send a message
Keep this terminal open.
Open another terminal:
python -m streamlit run app.py-
Select:
- Domain
- Difficulty Level
- Team Size
- Tech Stack
- Hackathon Theme
-
Click:
Generate Project
- AI Agents will:
- Generate project ideas
- Suggest technologies
- Create documentation
- Export generated output as Markdown.
HackForge AI uses multiple AI agents that collaborate together to generate structured hackathon-ready outputs.
Workflow:
User Input
↓
Idea Generator Agent
↓
Tech Stack Agent
↓
README Agent
↓
Final AI Output
✅ Local AI Processing
✅ No Paid API Required
✅ Lightweight AI Architecture
✅ Offline AI Support
✅ Portfolio-Ready Project
✅ Modern AI SaaS UI
hackforge-ai/
│
├── app.py
├── requirements.txt
├── README.md
├── Dockerfile
└── .dockerignore
- More AI Agents
- PDF Export
- Agent Memory
- Workflow Visualization
- Real-time Collaboration
- Cloud Deployment
- Authentication System
- Project Saving Feature
This project was built as part of the:
Build with AI – Agent Builder Camp
by GeeksforGeeks × Google for Developers.
The workshop helped in understanding:
- AI Agents
- Local LLM Workflows
- Prompt Engineering
- AI Product Building
- Streamlit UI Development
This project is licensed under the MIT License.
Passionate about:
- AI/ML
- Open Source
- Streamlit Apps
- Multi-Agent Systems
- Data Science


