llmledger

llmledger is a helper that answers a simple question: which AI model should I use for this, and what will it cost? You paste a prompt or describe a job, and it suggests models with prices, quality ratings and how likely their prices are to change. It is handy for anyone who pays for AI models and wants to pick sensibly.

What this MCP server does

You ask your AI something like which model to use for a coding task or what a daily workload would cost. Your AI passes the question to this helper. The helper checks llmledger's price catalogue, which tracks every price change since 2023, plus Arena quality ratings. It works out the best value, top rated and cheapest options, and sends the answer back to your chat. You see model names, prices and estimates in plain words.

What you can do with it

• Suggest a model and effort level for a project
• Compare up to five models side by side
• Estimate a budget for a daily workload
• Find models within your price limits
• Show a model's price history and outlook
• See which new models are not priced yet
• Estimate what your own coding agent usage would have cost on other models

Try asking your AI

• “Which model should I use for writing product descriptions, and what will 500 of them cost?”
• “Compare gpt-5, claude-sonnet-4-5 and deepseek-chat for coding quality and price.”
• “What is a reasonable monthly budget if I run 200 short prompts a day?”
• “Has the price of claude-sonnet-4-5 changed recently, and is it likely to change in the next 90 days?”

What it gives back to you

You get a short answer in the chat: model names, their prices per million tokens, a cost estimate for your prompt or workload, and a quality rating where available. It may also show how likely a price change is in the next 90 days. For comparisons, you get a small side-by-side list.

Before you start

Who it's for

People who choose and pay for AI models, like product managers, writers, developers and small teams watching their AI budget.

Jobs it suits