DEV Community

Lasya Lingam
Lasya Lingam

Posted on AI-assisted

CloudBlueprint AI - Hacktoberfest Open-Source AI Challenge: Week 2

#devchallenge #hf26challenge

Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle Submission 💸

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

CloudBlueprint AI turns an early-stage project idea into a practical AWS project plan. A user describes the problem, intended users, and desired features, and the app generates a structured plan with a project summary, suggested features, AWS architecture, data flow, security considerations, and build phases.
It’s aimed at indie developers, students, and small teams who want help turning an idea into an initial cloud design. It could save time on early planning and provide a useful starting point before implementation.

Demo

How CloudBlueprint AI works

Code

GitHub logo Meher01 / CloudBlueprint-AI

CloudBlueprint AI - Hacktoberfest Open-Source AI Challenge: Week 2

CloudBlueprint AI turns a project idea into a practical AWS project plan. Enter a project title, the problem it solves, intended users, and any desired features. The app checks that the idea is understandable, then generates a plan with an AWS architecture, data flow, security considerations, and build phases.

The project uses a React and TypeScript frontend, a FastAPI backend, and the Qwen 2.5 3B model running locally through Ollama. Generated plans are suggestions, not production-ready architecture designs.

Features

  • Generate a structured plan with six sections
    • Project summary
    • Suggested features
    • AWS architecture
    • Data flow
    • Security and privacy considerations
    • Build phases
  • Reject input that the language model identifies as unintelligible.
  • Display section headings in bold and list items as bullets.
  • Download a generated plan as a text file.
  • Run the language model locally with Ollama.

Technology

  • Frontend: React, TypeScript, and Vite
  • Backend…
View on GitHub

How I Built It

CloudBlueprint AI uses Qwen 2.5 3B, an open-weight model run locally with Ollama. The frontend is built with React and TypeScript, and the backend uses Python and FastAPI to send project details to the model and return its generated plan.
The model is central to the product: it interprets each project idea and creates a tailored plan, rather than returning a fixed template. The prompt guides it to use AWS services in the proposed architecture and to explain their roles, tradeoffs, and security considerations.

Why Does Open Innovation Matter?

Open-weight models made it possible to build and experiment with an AI-powered planning tool while running inference locally. That gave me control over which model to use and how it responds, without making the project depend on a closed model API.
It also makes the project easier to adapt: developers can inspect the model options, change the prompts, and run the system in their own environment.

Top comments (0)