Skip to content

ComparEdge LLM Cost MCP Server

Repository: comparedge/llm-cost-mcp By: ComparEdge Stars: 1 Category: Finance, AI Operations, Cost Management Auth: None (free, no API key) Pricing: Free

Overview

The ComparEdge LLM Cost MCP server gives AI agents real-time token cost calculations for 69 models across 17 providers. Operators deploying AI agents at scale need cost awareness — this tool answers "what will this API call cost?" before the call is made.

Prices are verified by ComparEdge, a SaaS pricing intelligence platform. The server covers OpenAI (GPT-4o, GPT-4.1, o3/o4), Anthropic (Claude Sonnet, Opus, Haiku), Google (Gemini), Meta (Llama), Mistral, DeepSeek, Cohere, xAI (Grok), and more.

What You Can Do

  • Calculate token costs for any supported model and provider
  • Compare costs across models for the same workload
  • Estimate API spend before running large batch jobs
  • Track per-model pricing changes over time
  • Integrate cost awareness into agent workflows

Installation

{
  "mcpServers": {
    "comparedge-llm-cost": {
      "command": "npx",
      "args": ["-y", "@comparedge/llm-cost-mcp"]
    }
  }
}

Or clone and run:

git clone https://github.com/comparedge/llm-cost-mcp.git
cd llm-cost-mcp
npm install
npm start

Use Cases for Business Operators

  1. Agent cost budgeting: Before deploying autonomous agents at scale, estimate per-task API costs
  2. Model selection: Compare GPT-4o vs Claude Sonnet vs Gemini for specific workloads by actual cost
  3. Vendor evaluation: Track which AI provider gives the best price-performance for your use case
  4. Spend forecasting: Project monthly API costs based on expected usage patterns
  5. Cost optimization: Identify overpriced models and switch to cheaper alternatives

Covered Providers (17)

OpenAI, Anthropic, Google (Gemini), Meta (Llama), Mistral, DeepSeek, Cohere, xAI (Grok), AI21 Labs, Amazon (Titan/Nova), Perplexity, Together AI, Fireworks AI, Replicate, Groq, Cerebras, SambaNova

Integration with CorpusIQ

Operators using CorpusIQ for business analytics can pair this with cost management workflows. As AI agent deployment grows, knowing per-task costs becomes as important as knowing per-transaction revenue.

Limitations

  • 1 GitHub star — brand new, community adoption pending
  • Prices updated by ComparEdge; accuracy depends on their verification cycle
  • Covers token costs only — doesn't account for infrastructure, latency, or throughput costs
  • No historical pricing data (yet — see ComparEdge Price Watch MCP for that)

Verdict

HIGH VALUE for any operator deploying AI agents at scale. The "no API key, free" model removes adoption friction. Particularly relevant now as enterprise AI spending moves from experimentation to production budgets. Complements the ComparEdge Price Watch MCP for a complete AI spend intelligence suite.

Discovered: July 14, 2026 via mcpservers.org. Guide by CorpusIQ MCP Discovery Engine.