ProjectMind AI
Agentic AI tool generating academic blueprints with Dual Mode (Local Ollama & Cloud Gemini API).
About the Project
An AI-powered tool that generates complete academic project blueprints for MCA / Engineering students. Originally designed to run fully on local LLMs using Ollama (100% offline, no paid APIs), the system was later enhanced with Google Gemini API support to enable fast and scalable cloud deployment.
How It Works
This project follows an Agentic AI pipeline with specialized agents:
User Input
Idea & Params
Planner Agent
Breaks down tasks
Generator Agent
Creates structured content
Formatter Agent
Cleans & organizes
Final Blueprint
PDF/DOCX Output
Dual Mode Support (Key Highlight)
This hybrid design ensures both privacy (local AI) and scalability (cloud AI).
- ✦Local Mode (Ollama): Runs fully on your system with no internet required and no API key needed. Uses models like Llama3 / Mistral.
- ✦Cloud Mode (Gemini API): Fast responses with no local setup required. Easily deployable on Streamlit Cloud.
Key Features
The system provides an end-to-end blueprint creation experience:
Generate full project blueprints and AI-powered structured outputs
Tech stack & dataset suggestions tailored to the project idea
Viva questions and Project timeline generation
Starter code suggestions
Export blueprints to PDF, DOCX, PPT, and Markdown
Tech Stack
Built entirely in Python, leveraging modern AI and Web frameworks:
Python & Streamlit (for the interactive frontend)
LangChain (for the Agentic AI pipeline orchestration)
Ollama (for local LLM inference)
Google Gemini API (for cloud inference)
FPDF / python-docx / python-pptx (for document generation)