Skip to content

GreenDev Coach

Scan your repo. Cut your carbon.

Sign in with your Google or GitHub account to get started.

The Problem

Green code is
good code.

The cloud is physical infrastructure that consumes real electricity. Every architectural choice a developer makes — which compute service, which region, how often CI runs — has a direct energy cost. Most developers never see that number. We think they should.

The numbers that matter

~1%

of global electricity

Data centres consume roughly 1% of global electricity — a share that grows every year as cloud adoption accelerates. The software choices developers make today determine the hardware utilisation of tomorrow.

40%

CI energy wasted

Studies show that 40% of CI compute time is spent re-downloading dependencies that haven't changed. A single actions/cache step eliminates that — for every project, on every run, forever.

14×

idle VM waste

An always-on EC2 instance runs 720 hours per month. A student project is actually used roughly 50 of those hours. The other 670 hours burn electricity for nothing. Serverless scales to zero.

95%

renewable in us-west-2

AWS Oregon (us-west-2) runs on more than 95% renewable energy. Moving a workload from ap-southeast-1 to us-west-2 can cut its effective carbon intensity by 40% with a single config change.

50 MB

vs 1.1 GB Docker image

node:20-alpine is 50 MB. node:latest is 1.1 GB. Every CI pull, every container start moves that delta across the wire. Alpine is not a trade-off — it's strictly better.

Our Principles

Visibility first

You cannot improve what you cannot measure. Most developers have no idea how many compute hours their project wastes per month, what that costs, or what it emits. GreenDev Coach makes that number concrete and personal.

Green is good engineering

Every green practice we recommend is also a performance or cost win. Alpine images build faster. Caching reduces pipeline time. Serverless eliminates idle spend. Sustainability and quality are the same thing.

Student devs are the leverage point

The habits a developer forms in university follow them to every startup, every enterprise, every cloud bill they'll ever write. Teaching green patterns early creates a compounding effect at industry scale.

Specific, not preachy

We don't show a generic "be sustainable" message. We show your repo, your CI triggers, your Docker base image, your region — and exactly what to change, with the YAML to paste directly into your workflow.

The Solution

How It Works

From repo URL to actionable sustainability report in under 30 seconds. Here's exactly what happens under the hood.

01

Paste your GitHub repo URL

Enter any public GitHub repository URL. GreenDev Coach uses the GitHub API to read your repo structure — no cloning, no credentials, no code execution.

  • Fetches your full file tree in under a second
  • Reads .github/workflows/, Dockerfile, and package.json
  • Detects languages, topics, and dependency counts
02

Configure your deployment

Tell us how your project is deployed. We use this to calculate real energy and cost baselines — an EC2 instance vs Lambda has a 14× difference in idle compute.

  • Cloud provider: AWS, GCP, Azure, or Auto-Detect
  • Compute type: VM, serverless, container, or static
  • Region: we know which regions run on renewable energy
  • CI/CD: frequency and tooling (GitHub Actions, CircleCI, etc.)
03

Five analysis engines run in parallel

GreenDev runs five specialised static analysis engines simultaneously. Each focuses on a different dimension of your project's environmental impact.

  • CI/CD Engine — detects trigger patterns, caching, path filters
  • Docker Engine — checks base images, multi-stage builds
  • Compute Engine — evaluates always-on vs serverless, instance sizing
  • Assets Engine — measures dependency footprint and image optimisation
  • Region Engine — maps your region to carbon intensity data
04

Sustainability score + 5 subscores

Your overall score (0–100) is calculated by deducting points for each detected issue. Five category subscores show exactly where to focus.

  • HIGH impact issues: −20 pts each
  • MEDIUM impact issues: −10 pts each
  • LOW impact issues: −5 pts each
  • Category grades: A, B, C, D, F per dimension
  • Before/after projection: CO₂ and cost savings
05

AI report in four formats

AWS Bedrock (Claude) generates four tailored versions of your report — each optimised for a different audience.

  • Plain English — for anyone, no jargon
  • Technical — for engineers and DevOps, implementation detail
  • Sustainability — carbon metrics and environmental context
  • Pitch-Ready — executive summary for stakeholders

The Five Analysis Engines

CI/CDBranch filters, path filters, caching, workflow count
DockerBase image weight, multi-stage builds, .dockerignore, root user
ComputeAlways-on VMs, instance sizing, serverless alternatives
AssetsDependency count, unoptimised images, bundle weight
RegionCarbon intensity tiers, renewable energy data, migration cost

Privacy & Security

We DO access

Public file tree

GitHub Actions YAML

Dockerfile

package.json

We NEVER access

Source code execution

Private repos

Secrets or env vars

Your AWS account

Built for the AWS Sustainability Challenge

About GreenDev

We built GreenDev Coach to empower developers to write software that actively respects our planet. A sustainability analysis tool that scans public GitHub repositories and gives student developers an actionable roadmap to greener, cheaper cloud deployments.

The best time to write green code was before you shipped. The second best time is now.

What We Built

1

Sustainability Engine

Static analysis across CI/CD, Docker, compute, assets

2

AI Report Generator

AWS Bedrock Claude — plain English to pitch-ready reports

3

Score & Subscores

Per-category grading across 5 dimensions

4

Framework Comparator

Green score benchmarks for Next.js, Astro, Svelte, Remix & more

Tech Stack

FrontendNext.js 16 + TypeScript
AIAWS Bedrock (Claude Sonnet)
DatabaseSupabase (scan history)
AnalysisGitHub API + 5 custom engines
HostingVercel Edge Network
Analyze Your Repository