mydentify
List product
Live
  1. Tired of managing Stripe invoices? just passed Don't get banned from Stripe for #5 on Best products for saving more money.
  2. Tired of managing Stripe invoices? is making a move.
  3. BuyMicroStartups just passed Mealime for #7 on Best products for saving more money.
  4. BuyMicroStartups just caught its first spark.
  5. Don't get banned from Stripe is making a move.
  6. Momentum is building behind Headroom.
  7. Mealime just passed Tyrodesk for #7 on Best products for saving more money.
  8. Mealime just caught its first spark.
  9. YNAB is making a move.
  10. Trendle is gaining speed.
Activity
Free ai model economics tool

Find What Drives Your AI Bill

Upload or paste AI usage records to total requests and model costs. CSV and JSON data stays in your browser.

No account requiredFiles stay in your browserReviewed August 1, 2026
Open the plain-text guide
Usage export

Add CSV or JSON usage

Recognized fields include model, input_tokens or prompt_tokens, output_tokens or completion_tokens, requests, cached_tokens, and cost. The file stays in your browser.

Your usage summary will appear here

Use provider totals when supplied and current catalog pricing when they are not.

How to use

How to analyze LLM usage costs

  1. 1

    Add the usage records

    Choose a CSV or JSON file, or paste rows directly.

  2. 2

    Normalize the token fields

    The analyzer recognizes common input, output, request, model, and cost columns.

  3. 3

    Review cost by model

    Find the models responsible for the most requests, tokens, and estimated spend.

Why this matters

See where AI spending goes

Provider totals can hide which model or workload created the bill. Normalized usage rows make costly traffic easier to find.

For editors and agents

Cite this tool accurately

Use the canonical page and this short description when you reference the tool. The plain-text guide keeps the method and limits easy to read.

LLM Usage Cost Analyzer — Upload or paste AI usage records to total requests and model costs. CSV and JSON data stays in your browser.

https://mydentify.com/tools/llm-usage-cost-analyzerRead the plain-text guide
Guide

Read an AI usage export

Use common provider usage fields

The analyzer accepts common prompt, completion, input, output, cached-token, request-count, and cost field names.

Prefer billed cost when available

A supplied cost field takes priority because it can reflect the provider's actual billing. Missing costs are estimated from the current public model catalog when the model ID matches.

Keep usage files in browser

Parsing and aggregation run in your browser. The page fetches public model prices but does not upload the usage file or request provider credentials.

Questions

Questions about LLM usage files

Which file formats are supported?

The first version accepts CSV and JSON usage rows.

Does the file leave my browser?

No. Parsing and aggregation run locally in the page.

Which column names are recognized?

Use model plus common input_tokens, prompt_tokens, output_tokens, completion_tokens, requests, cached_tokens, or cost fields.

What if my model is unknown?

Its token and request totals remain visible, but estimated cost stays at zero unless the row supplies a cost.

Why use the supplied cost first?

Provider totals can include exact tokenization and billing rules that a public list-price estimate cannot reproduce.

Can this connect to my provider account?

No. Version one intentionally uses exports and asks for no API credentials.

Keep working
All tools
ToolsAI Model Cost CalculatorEstimate and compare monthly AI model API costs.ToolsMRR CalculatorSee what is growing or shrinking your recurring revenue.
Back to all tools