Skip to main content
Glama
SOMA AI Emission Factors (SAEF)

SOMA AI Emission Factors (SAEF)

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
SAEF_REGIONNoOptional default region. The region comes from the `region` argument, then from the `SAEF_REGION` environment variable, then falls back to the global average.

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": false
}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}
experimental
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
estimate_footprintA

Estimate carbon, energy and water for AI inference using SOMA AI Emission Factors (SAEF).

Args: tokens: Billed tokens, input plus output, as shown on the provider invoice or usage page. model: Model identifier such as "gpt-4o", "claude-sonnet-4-5" or "gemini-2.5-flash". Unmapped models are treated as the conservative class C. region: Cloud region id (for example "eu-west-1", "us-east-1"), ISO-3 country code (for example "ESP"), eGRID subregion, "eu-27" or "global" for the world-average grid. Use list_regions to see options.

Returns carbon in kg and g CO2e with a central and a high case, facility energy in Wh, water in mL where the region has a water factor, the model class (A small, B mid-size, C large), the region used, the dataset version and a provenance trail to the source CSV rows. All figures are estimates, not measurements. They exclude model training development lines and are intended for Scope 3 category 1 reporting.

estimate_chat_usageA

Estimate carbon, energy and water for chat usage, counted in messages, using SOMA AI Emission Factors (SAEF).

Each message is one chat exchange (a question and its answer) of 1,210 tokens, the dataset's short-chat default of 928 input and 282 output tokens. Long documents, coding agents and reasoning modes use far more tokens, so use estimate_footprint when you know the token count.

Args: messages: Number of chat exchanges. model: Model identifier such as "gpt-4o" or "claude-sonnet-4-5". region: Cloud region id, ISO-3 country code, eGRID subregion, "eu-27" or "global". Use list_regions to see options.

Same output as estimate_footprint, plus the token assumption used. All figures are estimates, not measurements.

model_classA

Return the SAEF model class (A small, B mid-size, C large) for a model identifier.

SOMA AI Emission Factors (SAEF) assign factors by class, not by brand. A model that matches no known pattern gets the conservative class C. Embeddings, image, audio, video and moderation models are out of scope and return an error.

list_regionsA

List the region codes accepted by the other SAEF tools.

Returns the special values ("global", "eu-27"), the cloud regions of AWS, Azure and Google Cloud with names, and how to pass a country. Any ISO-3 country code (for example "ESP", "DEU", "USA") or eGRID subregion (for example "RFCE") also works.

compare_regionsA

Compare the carbon and water estimate for the same workload across regions using SOMA AI Emission Factors (SAEF).

Args: tokens: Billed tokens, input plus output. model: Model identifier such as "gpt-4o". regions: Region codes to compare, for example ["eu-north-1", "us-east-1", "ap-south-1"]. See list_regions.

Returns one row per region, sorted from lowest to highest carbon, with kg and g CO2e (central and high case) and water in mL. The lowest-carbon grid is not always the lowest-water one. A region that is not recognised gets an error entry instead of failing the whole call. Estimates, not measurements.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription
aboutAbout SAEF, its licence, citation and links.

TDQS

A4.3/5.0

Scored across 5 tools

Disambiguation5/5

Each tool targets a distinct input mode: token-based estimation, message-based chat estimation, multi-region comparison, region enumeration, and model classification. The descriptions explicitly route between the two estimate tools ("use estimate_footprint when you know the token count"), eliminating the one plausible overlap. list_regions and model_class are clearly non-overlapping lookups.

Naming Consistency4/5

Four of five tools follow a clean verb_noun snake_case pattern (estimate_footprint, estimate_chat_usage, list_regions, compare_regions). model_class is a bare noun and would read more consistently as get_model_class, a minor deviation in an otherwise predictable set.

Tool Count5/5

Five tools is well within a comfortable range and each one earns its place: two estimation entry points, one comparison, one reference lookup, one classifier. There is no filler or redundant wrapper.

Completeness4/5

The surface covers the core workflow (classify model, list regions, estimate by tokens or messages, compare regions) with no dead ends for the stated Scope 3 reporting use case. The notable gap is the absence of any helper to convert text or document content into billed tokens, so an agent holding raw text must guess the token count before calling estimate_footprint.

Maintenance

ActivityMaintained
ResponsivenessNo recent community issues