# GreenPick

> GreenPick is an MCP server for cloud and infrastructure. Carbon, cost and latency tradeoffs for cloud region and scheduling choices, in-editor. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Platforms:** CLI
- **Repository:** https://github.com/bithug01/greenpick
- **MCP Registry name:** io.github.bithug01/greenpick
- **Categories:** [cloud and infrastructure](https://ai-powered.com/categories/cloud-and-infrastructure)
- **Listing updated:** 2026-11-14

## About

GreenPick answers while you are picking where to deploy, when to run a job or how to right-size, rather than reporting emissions afterwards. Its README gives an example: the same 40-GPU-hour training run emits about 0.6 kg CO2 in Stockholm against about 17.5 kg in Sydney.

It covers AWS, GCP and Azure across 25 regions, for web services, GPU training and deferrable jobs. Live grid data comes from EIA-930 for the US and ENTSO-E for the EU (both optional keys), with Hydro-Quebec, IESO and TEPCO needing none. Without keys it falls back to labelled typical values built from a year of real data. Cost figures come from a static table, never invented by the model.

The README is blunt about scope: this is directional guidance for decisions, not audited carbon accounting or ESG reporting, and it does not claim GHG Protocol or ISO compliance. It also flags backfires, such as a greener region losing to egress cost or crossing a data residency line.

## Features

- **Tri-axis region ranking:** Ranks regions by carbon with cost and latency rated red, yellow or green.
- **Carbon-aware scheduling:** Finds the cleanest hours or time of day for deferrable jobs.
- **GPU and LLM carbon:** Estimates training and inference footprints with where and when advice.
- **Backfire guards:** Warns when egress or data residency undo a greener choice.
- **Day-ahead forecast:** Uses EU forecasts to pick upcoming clean windows.
- **Works offline:** Typical grid values apply when no API keys are set.

## Good for

- Choose a cloud region for a new service
- Time a nightly batch job for lower-carbon hours
- Estimate the carbon of a GPU training run
- Right-size workloads for cost and carbon together

## Install

### Claude Code (Terminal)

```sh
claude mcp add greenpick --env EIA_API_KEY='<EIA_API_KEY>' --env ENTSOE_TOKEN='<ENTSOE_TOKEN>' --env ELECTRICITYMAPS_TOKEN='<ELECTRICITYMAPS_TOKEN>' -- uvx greenpick
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "greenpick": {
      "command": "uvx",
      "args": [
        "greenpick"
      ],
      "env": {
        "EIA_API_KEY": "<EIA_API_KEY>",
        "ENTSOE_TOKEN": "<ENTSOE_TOKEN>",
        "ELECTRICITYMAPS_TOKEN": "<ELECTRICITYMAPS_TOKEN>"
      }
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "greenpick": {
      "command": "uvx",
      "args": [
        "greenpick"
      ],
      "env": {
        "EIA_API_KEY": "<EIA_API_KEY>",
        "ENTSOE_TOKEN": "<ENTSOE_TOKEN>",
        "ELECTRICITYMAPS_TOKEN": "<ELECTRICITYMAPS_TOKEN>"
      }
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "greenpick": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "greenpick"
      ],
      "env": {
        "EIA_API_KEY": "<EIA_API_KEY>",
        "ENTSOE_TOKEN": "<ENTSOE_TOKEN>",
        "ELECTRICITYMAPS_TOKEN": "<ELECTRICITYMAPS_TOKEN>"
      }
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "greenpick": {
      "command": "uvx",
      "args": [
        "greenpick"
      ],
      "env": {
        "EIA_API_KEY": "<EIA_API_KEY>",
        "ENTSOE_TOKEN": "<ENTSOE_TOKEN>",
        "ELECTRICITYMAPS_TOKEN": "<ELECTRICITYMAPS_TOKEN>"
      }
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "greenpick": {
      "command": "uvx",
      "args": [
        "greenpick"
      ],
      "env": {
        "EIA_API_KEY": "<EIA_API_KEY>",
        "ENTSOE_TOKEN": "<ENTSOE_TOKEN>",
        "ELECTRICITYMAPS_TOKEN": "<ELECTRICITYMAPS_TOKEN>"
      }
    }
  }
}
```

## Tools

- `compare_regions_tri_axis`: Ranks candidate regions by carbon with cost and latency tradeoffs.
- `greener_alternative_to`: Names the greenest region within a latency allowance of your pick.
- `region_carbon_now`: Current carbon intensity for a region with source and freshness.
- `cleanest_hours_to_run`: Cleanest recent hours to time a deferrable job.
- `cleanest_time_of_day`: Cleanest time of day for a recurring job.
- `greenest_regions_on_average`: Regions ranked on a typical day in a month or year.
- `carbon_forecast`: Cleanest upcoming hours from the EU day-ahead forecast.
- `estimate_workload_carbon`: Rough monthly kgCO2 for a workload in a region.
- `estimate_training_carbon`: Carbon of a GPU training run with where and when advice.

## FAQ

### Is it audited carbon accounting?

No. The README says it is directional guidance, not ESG or Scope 2 reporting.

### Does it need API keys?

No. Keys for EIA, ENTSO-E or Electricity Maps are optional.

### Which clouds does it cover?

AWS, GCP and Azure, 25 regions.

### How many tools does it have?

20.

### How do I add GreenPick to Claude Code?

Run this command in your terminal: claude mcp add greenpick --env EIA_API_KEY='<EIA_API_KEY>' --env ENTSOE_TOKEN='<ENTSOE_TOKEN>' --env ELECTRICITYMAPS_TOKEN='<ELECTRICITYMAPS_TOKEN>' -- uvx greenpick

### What are the best alternatives to GreenPick?

The closest alternatives to GreenPick are GridHop GPU Price & Carbon Index, GreenOps Scan, Tidegap, BoaviztAPI and CloudGPU MCP.

## Where this comes from

- Description: written from the project's README.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [GridHop GPU Price & Carbon Index](https://ai-powered.com/mcp/gridhop-gpu-price-and-carbon-index): GridHop GPU Price & Carbon Index is an open-source MCP server for cloud and infrastructure.
- [GreenOps Scan](https://ai-powered.com/mcp/greenops-scan): GreenOps Scan is a free MCP server for cloud and infrastructure.
- [Tidegap](https://ai-powered.com/mcp/tidegap): Tidegap is an open-source MCP server for ai models and llm tooling.
- [BoaviztAPI](https://ai-powered.com/mcp/boaviztapi): BoaviztAPI is an open-source MCP server for cloud and infrastructure.
- [CloudGPU MCP](https://ai-powered.com/mcp/cloudgpu-mcp-2): CloudGPU MCP is a paid MCP server for cloud and infrastructure.
- [GPU Finder](https://ai-powered.com/mcp/gpu-finder): GPU Finder is a freemium MCP server for cloud and infrastructure.
- [gpu.nz GPU cloud](https://ai-powered.com/mcp/gpu-nz-gpu-cloud): gpu.nz GPU cloud is a paid MCP server for cloud and infrastructure.
- [CloudCostAI](https://ai-powered.com/mcp/cloudcostai): CloudCostAI is a freemium MCP server for cloud and infrastructure.

---
Source: https://ai-powered.com/mcp/greenpick
