REST API v2

AI Impact Calculator API

Calculate the carbon, energy, and region-aware water footprint of any AI model call. The only API with water stress scoring per data center region.

v2 adds 14 data center regions with WUE data, 13 AI models, input/output token split, and water stress classification from WRI Aqueduct.

Get API Key on RapidAPI

What's New in v2

💧

Region-Aware Water Impact

14 data center regions with specific WUE and water stress classification from WRI Aqueduct.

🤖

13 AI Models

GPT-4o, Claude 3.5 Sonnet, Gemini 1.5, Mistral, Llama 3, and more.

⚡

Input/Output Token Split

Output tokens use ~3x more energy. Pass them separately for accurate estimates.

🔄

Backwards Compatible

Old v1 grid keys still work. v1 endpoint remains unchanged.

Base URL

POST https://aiimpactcalculator.com/api/v2/calculate

GET  https://aiimpactcalculator.com/api/v2/calculate  (list models & regions)

Authentication

Pass your API key as a Bearer token in the Authorization header.

Authorization: Bearer YOUR_API_KEY

🔑

Get your API key for free

Subscribe on RapidAPI - free plan includes 100 requests/month, no credit card required. Subscribe now →

Endpoint

POST/api/v2/calculate

Request Body

Field	Type	Required	Description
model	string	Yes	AI model key (see table below)
tokens	number	Yes*	Total tokens. *Or provide input_tokens + output_tokens instead.
input_tokens	number	No	Input/prompt tokens (for more accurate estimates)
output_tokens	number	No	Output/completion tokens (uses ~3x more energy)
region	string	No	Data center region key (default: `global-avg`)
custom_wh_per_1k	number	Conditional	Required when model is `custom`

Models

model	Provider	Description	Energy rate
gpt-4o	OpenAI	GPT-4o	3.5 Wh / 1k tokens
gpt-4	OpenAI	GPT-4 / GPT-4 Turbo	4.5 Wh / 1k tokens
gpt-3.5	OpenAI	GPT-3.5 Turbo	0.5 Wh / 1k tokens
claude-3.5-sonnet	Anthropic	Claude 3.5 Sonnet	2.0 Wh / 1k tokens
claude-3-opus	Anthropic	Claude 3 Opus	4.5 Wh / 1k tokens
claude-3-haiku	Anthropic	Claude 3 Haiku	0.3 Wh / 1k tokens
gemini-1.5-pro	Google	Gemini 1.5 Pro	3.0 Wh / 1k tokens
gemini-1.5-flash	Google	Gemini 1.5 Flash	0.4 Wh / 1k tokens
llama-70b	Meta	Llama 3 70B	0.5 Wh / 1k tokens
llama-7b	Meta	Small Models (7B)	0.1 Wh / 1k tokens
mistral-large	Mistral AI	Mistral Large	2.5 Wh / 1k tokens
image-gen	Various	Image Generation	2.9 Wh / image
custom	-	Custom	Provide custom_wh_per_1k

RegionsNEW in v2

Each region has its own carbon intensity, Water Usage Effectiveness (WUE), and water stress level from the WRI Aqueduct Water Risk Atlas.

region	Location	Carbon (gCO2/kWh)	WUE (L/kWh)	Water Stress
us-east-virginia	US East (Virginia)	310	1.8	high
us-west-oregon	US West (Oregon)	80	0.8	moderate
us-west-california	US West (California)	210	1.2	critical
us-central-iowa	US Central (Iowa)	420	1.5	moderate
us-central-texas	US Central (Texas)	350	2.0	high
eu-ireland	EU (Ireland)	300	0.9	low
eu-netherlands	EU (Netherlands)	330	1.1	moderate
eu-germany	EU (Germany)	350	1.4	moderate
eu-sweden	EU (Sweden)	20	0.3	low
eu-france	EU (France)	55	0.6	low
asia-singapore	Asia (Singapore)	410	2.2	high
asia-japan-tokyo	Asia (Tokyo)	450	1.6	moderate
me-uae	Middle East (UAE)	500	3.0	critical
global-avg	Global Average	475	1.8	moderate

Example Response

{
  "model": "gpt-4o",
  "input_tokens": 30000,
  "output_tokens": 20000,
  "tokens": 50000,
  "region": "us-east-virginia",
  "region_label": "US East (Virginia)",
  "energy_kwh": 0.2825,
  "carbon_gco2e": 87.575,
  "water": {
    "liters": 0.5085,
    "stress_level": "high",
    "wue_liters_per_kwh": 1.8,
    "note": "US East (Virginia) is classified as high water stress by WRI Aqueduct. Evaporative cooling places significant pressure on local water resources."
  },
  "equivalents": {
    "smartphones_charged": 23.54,
    "km_driven": 0.515,
    "led_hours": 31.39,
    "netflix_hours_sd": 3.67,
    "google_searches": 942,
    "water_bottles_500ml": 1.02,
    "emails_equivalent": 22
  },
  "data_sources": [
    "EIA State Electricity Profiles (2025)",
    "WRI Aqueduct Water Risk Atlas"
  ],
  "plan": "free",
  "methodology_url": "https://aiimpactcalculator.com/methodology"
}

💧 Understanding Water Stress

Water Usage Effectiveness (WUE) measures how many liters of water a data center consumes per kWh of IT energy. Higher WUE means more water consumption. The water stress level adds geographic context from the WRI Aqueduct Water Risk Atlas.

low Low Stress

Abundant water supply. Cool climate reduces cooling needs. Example: Sweden, France.

moderate Moderate Stress

Water consumption is notable but manageable relative to supply. Example: Oregon, Netherlands.

high High Stress

Evaporative cooling places significant pressure on local water. Example: Virginia, Texas.

critical Critical Stress

Data centers compete directly with agriculture and municipal needs. Example: California, UAE.

Code Snippets

curl -X POST 'https://aiimpactcalculator.com/api/v2/calculate' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -d '{
    "model": "gpt-4o",
    "tokens": 50000,
    "region": "us-east-virginia"
  }'

Error Codes

HTTP Status	code	Meaning
401	UNAUTHORIZED	Missing or invalid API key
400	INVALID_MODEL	Unknown model value
400	INVALID_TOKENS	tokens must be positive; or provide input_tokens + output_tokens
400	MISSING_CUSTOM_RATE	custom_wh_per_1k required for custom model
400	BAD_REQUEST	Malformed JSON body
429	RATE_LIMIT_EXCEEDED	Monthly request limit reached

Pricing

Plan	Price	Requests/mo	Best for
Basic	$0 / mo	100	Evaluation, personal projects
Pro	$9 / mo	5,000	Solo devs, small apps
Ultra	$29 / mo	50,000	SaaS products, agencies
Mega	$99 / mo	500,000	Enterprise, ESG reporting

View plans and subscribe on RapidAPI →

Ready to integrate?

The only AI emissions API with region-aware water stress scoring. Free plan available.

Get API Key on RapidAPI Read the Methodology