Skip to content
MCP server

BoaviztAPI

Estimate the carbon footprint of cloud instances, servers, GPUs and devices

Runs locally
Install
Pricing
Open source
Platforms
CLI, self-hosted, Linux
Runs
Locally
Version
0.1.2
Listing updated
5 days ago

Similar MCP servers

About BoaviztAPI

BoaviztAPI is an open-source MCP server for cloud and infrastructure. It runs locally on your machine and is started by your AI client.

Footprint figures come from BoaviztAPI, built by Boavizta, and the server reports them with units, uncertainty ranges and caveats. It is an unofficial community project. Embedded impact (manufacturing) and use-phase impact (electricity) are always reported separately, with a total only when both exist, and a figure that BoaviztAPI cannot supply is shown as not available rather than zero. Defaults cover climate change (gwp), mineral and fossil resource use (adp) and primary energy (pe).

Eight read-only tools handle the work: list_impact_criteria, search_catalog, describe_catalog_item, estimate_cloud_instance, estimate_server, estimate_component, estimate_device and compare_estimates, which ranks two to five options and says when ranges overlap. Names are matched carefully, so a vague name such as h100 returns candidates instead of a silent guess.

It runs locally over stdio via uvx, Docker or from source and needs uv. By default it queries the public BoaviztAPI demo instance; the README recommends running your own with the Boavizta Docker image and setting BOAVIZTAPI_URL for regular use. Responses are cached for a day by default.

Install

Run in a terminal

claude mcp add boaviztapi-mcp -- uvx boaviztapi-mcp

Add to claude_desktop_config.json

{
  "mcpServers": {
    "boaviztapi-mcp": {
      "command": "uvx",
      "args": [
        "boaviztapi-mcp"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "boaviztapi-mcp": {
      "command": "uvx",
      "args": [
        "boaviztapi-mcp"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "boaviztapi-mcp": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "boaviztapi-mcp"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "boaviztapi-mcp": {
      "command": "uvx",
      "args": [
        "boaviztapi-mcp"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "boaviztapi-mcp": {
      "command": "uvx",
      "args": [
        "boaviztapi-mcp"
      ]
    }
  }
}

Using several servers? Build one config file with the MCP config generator.

Configuration

  • BOAVIZTAPI_URLBoaviztAPI base URL; use your own instance for regular use
  • BOAVIZTAPI_TIMEOUTSeconds to wait for each BoaviztAPI request
  • BOAVIZTAPI_DEFAULT_CRITERIAComma-separated impact criteria used when a tool call names none
  • BOAVIZTAPI_CACHE_TTLSeconds to cache list and lookup responses; 0 disables caching

Tools this server provides (8)

Taken from the project documentation; the running server may differ.

list_impact_criteria
Explains impact criteria codes and their units.
search_catalog
Finds exact catalogue names before estimating.
describe_catalog_item
Shows what an instance, archetype, GPU or CPU contains.
estimate_cloud_instance
Estimates the footprint of one cloud VM.
estimate_server
Estimates the footprint of an on-premise server, optionally with GPUs.
estimate_component
Estimates the footprint of one GPU, CPU, RAM module, SSD or HDD.
estimate_device
Estimates the footprint of a laptop, phone, monitor or other device.
compare_estimates
Ranks two to five options over the same period.

What it does

Cloud instance estimates
Footprint of one cloud VM by provider, instance type, region and load.
Server and GPU estimates
Estimate an on-premise server, optionally with GPUs, or a single component.
Device estimates
Cover laptops, phones, monitors, IoT devices and other end-user hardware.
Side-by-side comparison
Rank two to five options over the same period and flag overlapping ranges.
Careful name matching
Ambiguous names are refused with a candidate list instead of being guessed.
Self-hostable backend
Point it at your own BoaviztAPI instance and tune timeout, criteria and caching.

Good for

  • Estimate the footprint of 4 H100 GPUs running a year in Germany
  • Compare an AWS instance across regions
  • Find the embedded carbon of a typical laptop
  • Rank two instance types by climate impact
  • Look up what an impact criterion such as adp means

Pricing

Open source

BoaviztAPI FAQ

Where do the numbers come from?

Every number comes from BoaviztAPI by Boavizta; the server is an unofficial community project.

Can I change something with it?

No. All tools are read-only.

Which instance does it query?

The public BoaviztAPI demo by default; set BOAVIZTAPI_URL to use your own.

How do I install it?

Run it with uvx boaviztapi-mcp, with Docker, or from source; uv is required.

Are results cached?

Lists and lookups are cached for 86400 seconds by default; set the TTL to 0 to disable.

Is BoaviztAPI free?

Yes. BoaviztAPI is open source, so you can use and self-host it for free.

How do I add BoaviztAPI to Claude Code?

Run this command in your terminal: claude mcp add boaviztapi-mcp -- uvx boaviztapi-mcp

What are the best alternatives to BoaviztAPI?

The closest alternatives to BoaviztAPI are GridHop GPU Price & Carbon Index, TheAIMeters, GreenPick, AI Footprint and CloudCostAI.

Where this comes from

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

How AIPoweredBot checks servers