SkyDeploy Companion

A conversational assistant that creates and manages AWS resources using simple natural language commands.

Comment

Inspiration

I wanted to explore how cloud deployments could feel simpler and more conversational, especially for beginners who often find the AWS Console intimidating. The idea of turning natural language into cloud actions seemed like a great way to make DevOps more accessible without relying on heavy AI agents. This inspired me to build a lightweight assistant that can understand basic commands and help automate small AWS tasks.

What it does

SkyDeploy Companion allows users to create and manage simple AWS resources using plain text instructions. You can ask it to create S3 buckets, deploy basic Lambda functions, or prepare a bucket for static website hosting. The assistant interprets the user message using a rule-based logic system and executes the required AWS operations through a clean command workflow displayed in a chat-style interface.

How we built it

The backend is built using Python, Flask, and Boto3. I wrote a small intent-detection system using regular expressions and keyword patterns to interpret natural language and map it to AWS actions. The frontend is a simple HTML, CSS, and JavaScript interface designed as a conversational chat box. It sends user messages to the Flask API and displays the system responses in real time. The combination keeps the architecture simple, lightweight, and easy to test.

Challenges we ran into

One of the biggest challenges was balancing flexibility in natural language with the reliability of rule-based parsing. Users phrase the same request in many different ways, so I had to design patterns that capture intent without misinterpreting commands. Another challenge involved AWS permissions, particularly around S3 website hosting and Lambda invocation roles. Troubleshooting these configurations helped solidify my understanding of IAM policies and resource permissions.

Accomplishments that we're proud of

I’m proud of building a working conversational interface that actually performs real cloud actions reliably without needing large AI models. Creating a minimal system that still feels interactive and helpful was a big milestone. I’m also proud that the entire setup works smoothly with a small codebase and remains easy for others to understand and extend.

What we learned

This project helped me understand:

How natural-language parsing can be built without full AI models

The workflow of AWS S3, Lambda, and IAM

How to design clean request-response flows using Flask

The importance of proper AWS permissions and safe error handling

How small automations can greatly simplify cloud operations

What's next for SkyDeploy Companion

Next, I plan to:

Add support for more AWS services like DynamoDB or CloudWatch

Improve the language interpreter so it understands more flexible phrasing

Add a resource dashboard to visualize created assets

Introduce optional integration with lightweight LLM APIs for better intent detection

Built With

Share this project:

Updates