How to Use the AI Cost Calculator API — Free REST + MCP Server

#ai #api #openai #mcp

"How much will this GPT-4o call cost?" If you've asked this while building an AI app, this API is for you. It calculates token costs across Claude, GPT-4o, Gemini, Llama, and Mistral models — with real pricing data updated weekly.

Try It Right Now

curl -s -X POST https://api.lazy-mac.com/ai-cost/api/v1/calculate \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-4o", "input_tokens": 1000, "output_tokens": 500}' | jq

Response:

{
  "model": "gpt-4o",
  "provider": "openai",
  "input_tokens": 1000,
  "output_tokens": 500,
  "input_cost_usd": 0.0025,
  "output_cost_usd": 0.005,
  "total_cost_usd": 0.0075,
  "rates": {
    "input_per_mtok": 2.5,
    "output_per_mtok": 10
  }
}

All Endpoints

Endpoint	Description
`POST /api/v1/calculate`	Calculate cost for a specific model
`POST /api/v1/estimate`	Estimate from text length
`POST /api/v1/compare`	Compare costs across models
`GET /api/v1/cheapest?input_tokens=N&output_tokens=N`	Find cheapest model
`GET /api/v1/models`	List all 14 supported models
`GET /api/v1/models/:provider`	Models by provider

Find the Cheapest Model

curl -s "https://api.lazy-mac.com/ai-cost/api/v1/cheapest?input_tokens=10000&output_tokens=5000" | jq

Compare Claude vs GPT-4o

curl -s -X POST https://api.lazy-mac.com/ai-cost/api/v1/compare \
  -H "Content-Type: application/json" \
  -d '{"models": ["claude-opus-4", "gpt-4o"], "input_tokens": 10000, "output_tokens": 2000}' | jq

Use as an MCP Server

Ask Claude: "How much would 10K tokens cost on GPT-4o vs Claude Sonnet?"

{
  "mcpServers": {
    "ai-cost": {
      "url": "https://api.lazy-mac.com/ai-cost/mcp"
    }
  }
}

Supported Models (14 total)

Anthropic (Claude Opus 4, Sonnet 4, Haiku 3.5), OpenAI (GPT-4o, GPT-4o-mini, o1, o3-mini), Google (Gemini 2.0 Flash, Gemini 1.5 Pro), Meta (Llama 3.3 70B), Mistral (Large, Small).