Skip to content

πŸ’‘ AI-powered FinOps Agent built with AWS Bedrock, FastAPI, and React. Automates cost analysis, optimization, and recommendations for MSPs and cloud teams.

Notifications You must be signed in to change notification settings

MrigankJaiswal-hub/finops-Agent

Repository files navigation

🧠 FinOps AI Agent

Smart, Automated Financial Operations Insights powered by AWS Bedrock + FastAPI


🌟 About the Project

FinOps AI Agent is an intelligent cloud-native assistant that helps organizations analyze, optimize, and automate financial operations (FinOps).

It combines the power of AWS Bedrock, FastAPI, and React to deliver real-time cost insights, AI-driven recommendations, and automated execution workflows β€” empowering IT, finance, and operations teams to make smarter, data-driven decisions with minimal manual effort.

Developed as part of advanced AI and cloud hackathons, the project demonstrates how agentic AI can transform traditional FinOps into an autonomous decision-making system.


βš™οΈ Key Features

Category Description
πŸ” Analyze Ingest and parse cloud billing CSV data from S3 to identify cost anomalies and waste.
πŸ’‘ Recommend Generate contextual AI-driven recommendations for optimization using Bedrock (Claude 3 Sonnet).
βš™οΈ Execute Trigger automation actions like instance right-sizing, budget alerts, and forecast updates.
πŸ“ˆ Track Monitor and visualize KPIs such as monthly trends, spend per service, and budget adherence.
πŸ” Secure Auth Integrated AWS Cognito authentication with OIDC and JWT token verification.
☁️ Cloud-Native Fully serverless backend with API Gateway, Lambda, S3, and DynamoDB.
πŸ“œ History Log Persist historical FinOps insights and user-triggered actions in DynamoDB.

🧩 Architecture Overview


[Frontend: React + OIDC]
|
↓
[API Gateway + Lambda (FastAPI via Mangum)]
|
↓
[Bedrock (Claude 3 Sonnet) + DynamoDB + S3]


🧠 Tech Stack

Layer Technologies
Frontend React (Vite) + TailwindCSS + react-oidc-context
Backend FastAPI + Mangum + SQLAlchemy
AI Integration AWS Bedrock (Anthropic Claude 3 Sonnet)
Storage S3 (billing data) + DynamoDB (history & logs)
Auth AWS Cognito OIDC
Infra & CI/CD AWS SAM + GitHub Actions
Deployment API Gateway + Lambda + CloudFront

πŸš€ Getting Started

🧱 Prerequisites

  • Node.js β‰₯ 18
  • Python β‰₯ 3.12
  • AWS CLI configured with SSO or IAM profile
  • AWS SAM CLI installed

πŸ”§ Setup Instructions

Clone the Repository

git clone https://github.com/MrigankJaiswal-hub/finops-Agent.git
cd finops-agent

Backend Setup

cd backend
python -m venv .venv
.venv\Scripts\activate  # (for Windows)
pip install -r requirements.txt
uvicorn app:app --reload

Frontend Setup

cd frontend
npm install
npm run dev

Visit http://localhost:5173


☁️ AWS Deployment via SAM

cd sam-deploy/finops-backend
sam build
sam deploy --guided \
  --profile 390503781686_AdministratorAccess \
  --region us-east-1 \
  --parameter-overrides \
    S3DataBucket=finops-demo-bucket3905 \
    ExistingActionsTableName=FinOpsActions

After deployment, note the API URL and Lambda function name displayed in the outputs.


πŸ—‚οΈ Repository Structure

finops-agent/
β”‚
β”œβ”€β”€ frontend/               # React + OIDC user interface
β”œβ”€β”€ backend/                # FastAPI app + AI logic + S3 & DynamoDB utils
β”œβ”€β”€ sam-deploy/             # AWS SAM templates for Lambda + API Gateway
β”œβ”€β”€ .gitignore
β”œβ”€β”€ README.md
└── requirements.txt

🧾 Release Notes

πŸ”Έ v1.0.0 – Initial Public Release

Date: November 2025 Highlights:

  • Core FinOps analysis & AI insights pipeline
  • Bedrock (Claude 3 Sonnet) integration
  • DynamoDB historical tracking
  • Secure Cognito authentication (OIDC)
  • Full stack AWS-native deployment

πŸ”Ή v1.1.0 – Upcoming

Planned Features:

  • Multi-tenant organization support
  • Enhanced FinOps forecasting with RAG pipeline
  • PDF-based executive summary generation
  • Multi-region Bedrock agent orchestration
  • Interactive FinOps dashboard & insights viewer

πŸ‘©β€πŸ’» Contributing

We welcome community contributions to make FinOps AI Agent even better!

  1. Fork the repository

  2. Create a new branch:

    git checkout -b feature/new-feature
  3. Commit your changes:

    git commit -m "Add new feature"
  4. Push and open a Pull Request

All contributions should follow:

  • PEP8 (Python backend)
  • ESLint + Prettier (frontend)
  • Include meaningful commit messages and documentation updates.

πŸ“§ Contact

Author: Mrigank Jaiswal


πŸ›‘οΈ License

MIT License Β© 2025 β€” FinOps AI Agent Built for the SuperOps | AWS | Google Cloud | AI Hackathons