Kaytu enables engineering, DevOps, and SRE teams to reduce cloud costs by recommending optimal workload configurations based on actual-usage, ensuring savings without compromise.

Overview

• Ease of use: One-line command. Use without modifying workloads or making configuration changes.
• Base on actual Usage: Analyzes the past seven days of usage from Cloud native monitoring (CloudWatch), including advanced AWS CloudWatch metrics (where available).
• Customize: Optimize for region, CPU, memory, network performance, storage, licenses, and more to match your specific requirements.
• Secure - no credentials to share; extracts required metrics from the client side
• Open-core philosophy Use without fear of lock-in. The CLI is open-sourced, and the Server side will be open-sourced soon.
• Coming Soon: Non-Interactive mode, Azure support, GPU Optimization, Credit utilization for Burst instances, and Observability data from Prometheus

Getting Started

1. Install Kaytu CLI

MacOS

brew tap kaytu-io/cli-tap && brew install kaytu

Linux

curl -fsSL https://raw.githubusercontent.com/kaytu-io/kaytu/main/scripts/install.sh | sh

Windows (and all Binaries) Download Windows (Linux, and MacOS) binary from releases

2. Login to AWS CLI

Kaytu works with your existing AWS CLI profile (read-only access required) to gather metrics.

To confirm your AWS CLI login is working correctly:

aws sts get-caller-identity

Click here to see how to log in to AWS CLI.

We respect your privacy. Our open-source code guarantees that we never collect sensitive information like AWS credentials, IPs, tags, etc.

3. Run Kaytu CLI

kaytu