NetGreener — profiler for AI infrastructure efficiency

AI coding agents help you write AI code. NetGreener helps you run it efficiently.

Measure GPU cost, cloud bill, and infrastructure efficiency before and after deployment. Carbon, water, and energy are included so you can report the bonus layer — not the lead.

Most coding tools help software work better. NetGreener helps you find the expensive part of an AI project, forecast cost before a run, and verify the result in CI/CD and production.

One continuous system

Stage	Where	What NetGreener does
1. Developer Agent	VS Code, Cursor, CLI, SDK	Explains how coding patterns, algorithms, libraries, models, and configurations affect CPU/GPU, memory, cost, energy, carbon, and water; recommends lower-impact alternatives; and guides junior and AI-assisted developers as they code.
2. Pre-Run and CI	Before merge, test, or release	Forecasts cost, resource demand, energy, carbon, and water; compares expected impact with a baseline; and applies efficiency checks before inefficient changes scale.
3. Runtime Control	Local, staging, cloud, production	Measures actual CPU/GPU, RAM, cost, energy, carbon, and water; then attributes impact to the responsible code change, release, model, service, pipeline, project, and team.
4. Learning Loop	Every release	Compares predicted and actual impact; validates recommendations; returns evidence to developers; and improves future forecasts, guidance, and infrastructure decisions.

What this extension does

In the editor, NetGreener connects you to that continuous system:

Resource-aware guidance — understand the resource consequences of coding and architecture decisions as you write
Live run metrics — CPU, GPU, memory, energy, cost, and estimated CO₂ / water for measured runs, shown in the sidebar
Code issue detection — static analysis that flags resource-inefficient patterns (redundant loops, excessive allocations, repeated I/O, and more) directly in your editor
AI-powered optimization — select any flagged function and get a rewritten, lower-impact version with an explanation
Estimates & insights — project-level forecasts and trends across your projects

Requirements

The extension requires the NetGreener CLI (netgreener) to measure and analyze your code.

If the CLI is not installed when the extension activates, you will be prompted automatically. You can also run NetGreener: Install or Update CLI from the Command Palette.

Install Globally with pipx — recommended; installs one shared CLI and keeps NetGreener dependencies isolated from your project
Install in Project Python Environment — advanced; may upgrade/downgrade project packages such as TensorFlow
Install with Current Python — advanced; uses python3/py -3 from the integrated terminal environment

To install manually:

python3 -m pip install --user --upgrade pipx
python3 -m pipx ensurepath
python3 -m pipx install --force netgreener

Getting started

Install the extension and open a Python / AI project folder
On first open, choose Scan This Project (or NetGreener: Analyze Project later)
Log in if asked — GitHub or email/password (NetGreener: Log In)
First scan takes about two minutes; issues appear in the NetGreener sidebar
Optional: Ctrl+Shift+G R to run and measure energy and GPU cost in the terminal

Features

Run Metrics

See energy consumption (Wh), CPU/GPU utilization, memory, estimated cost, CO₂, and related resource metrics for every run — updated automatically when a run completes.

Code Issues

Scan your project for resource-inefficiency patterns with NetGreener: Scan Project for Issues ($(search) button in the sidebar). Issues are shown as:

Inline editor diagnostics with severity indicators
A filterable tree view in the sidebar
CodeLens actions on flagged functions

AI-Powered Optimization

Click Optimize Function on any flagged issue to send it to the NetGreener AI agent and receive a rewritten, lower-impact version. Supports applying fixes one-by-one or all at once with Optimize All Issues.

Estimates

The Estimates panel projects expected resource use, cost, and environmental impact based on historical runs and detected code patterns.

Why NetGreener is different

General-purpose coding agents primarily help teams generate, correct, and accelerate software. NetGreener adds a different engineering objective: understanding and reducing what software consumes. Faster code is not always more resource-efficient; NetGreener helps teams optimize performance, resource use, infrastructure cost, and environmental impact together.

Configuration

Setting	Default	Description
`netgreener.cliPath`	`netgreener`	Path to the CLI executable
`netgreener.apiBaseUrl`	(cloud)	NetGreener API base URL
`netgreener.llmProvider`	(server default)	AI provider: `azure_openai`, `openai`, `anthropic`, `ollama`, `custom`
`netgreener.llmKey`		API key for your AI provider
`netgreener.llmEndpoint`		Endpoint URL (required for hosted providers, Ollama, or custom)
`netgreener.llmModel`		Model or deployment name
`netgreener.dataPath`		Path or URI to training/inference data; required for full CNN/data feature extraction on AI projects (supports local paths, Azure Blob, S3, GCS)

Bring your own AI provider

NetGreener supports any AI provider for code optimization. Set llmProvider + llmModel (and llmKey / llmEndpoint as needed) in your VS Code settings. Supported providers:

openai — OpenAI-compatible hosted API
anthropic — Anthropic API
azure_openai — Azure-hosted deployment
ollama — local Ollama instance (http://localhost:11434)
custom — any OpenAI-compatible endpoint

Keyboard shortcuts

Shortcut	Action
`Ctrl+Shift+G R`	Run script in integrated terminal
`Ctrl+Shift+G L`	Open last run in dashboard
`Ctrl+Shift+G P`	Open project in dashboard
`Ctrl+Shift+G S`	Select active project

Commands

All commands are available from the Command Palette (Ctrl+Shift+P):

NetGreener: Log In / Log In with GitHub / Log Out
NetGreener: Run in Integrated Terminal
NetGreener: Scan Project for Issues
NetGreener: Optimize Function
NetGreener: Optimize All Issues
NetGreener: Open Project in Dashboard
NetGreener: Open Last Run in Dashboard
NetGreener: Select Project
NetGreener: Create Project for This Workspace

Privacy & data

Energy measurements and code analysis results are sent to the NetGreener platform associated with your account. Code sent for AI-powered optimization goes to your configured AI provider (or the NetGreener server default). No data is shared with third parties.

Support

Make hidden compute, energy, and environmental costs visible before code scales; then prove the improvement at runtime.

Website: netgreener.ai
Email: info@netgreener.ai