AI-Based Infrastructure Analysis

A powerful, automated report-generation tool for Sales Engineers, leveraging real cost data (InfraCost) and AI-driven reasoning (Cohere) to produce PDF documents highlighting how Weka can optimize AI training and inference workflows.

Weka AI Analysis & Report Generator

This tool collects your current infrastructure details—like your cloud provider, GPU usage, data size, and custom cost parameters—to compute existing spend and forecast how Weka's solution can reduce those costs. It automatically injects real or approximate pricing data from InfraCost and uses Cohere to tailor the final PDF analysis, bridging numeric cost formulas with reasoned AI commentary.

Quick Tip: Enter your environment info (AWS, GCP, etc.) along with the relevant cost parameters (like cloud download fees or idle GPU hours). The system will compile a professional LaTeX PDF describing how Weka fits your unique scenario.

1. Cloud Environment


2. Weka Cost Parameters

The fields below represent key cost values used in generating the final cost comparisons. Adjust them as needed for your environment's scale, e.g., approximate cost to download data, egress fees, or the reduced idle time Weka can enable.


3. Use Local Analysis (Optional)

To deepen the accuracy of your TCO calculations, you can run a local GPU analysis script on one of your own servers. This script collects data about GPU utilization, NFS usage, and more, then securely sends that info to our pod-efficiency.tools API for integration into your final PDF analysis.

Note: You only need to run this if you want a more precise real-time snapshot of your environment. Otherwise, you can rely on the estimates above.

wget https://api.pod-efficiency.tools/sbin/gpu-analyzer.py | python3 gpu-analyzer.py

Copy this command and run it on a local system with Python 3 installed. The script will gather GPU & NFS stats and automatically pipe them to https://api.pod-efficiency.tools/ in JSON format.


© 2025 WEKA IO LTD | AI-based Infrastructure Analysis Tool v0.3.1