Free AI cost tools

Analyze bills.
Compare AI costs.

For builders and small teams using several AI tools: understand what you pay, see what is missing, and choose one change to test.

Free tools · no account or payment needed for browser analysis · exports stay in browser memory

Try the example → bring a supported export → review your costs → download your analysis. Formats & first steps →

Browser tools

Analyze your costs in the browser

Start with the question you have today. Try sample data before bringing your own.

Model pricing

Compare model prices

Explore model prices, enter your expected usage, and compare estimated input, output and cache costs. Inspect the source and observation date behind each figure.

Compare model prices →

OpenRouter channel prices. A lower estimate needs a quality and reliability test on your own work.

Billing exports

Analyze your AI bill

Bring API usage exports together, add subscriptions, and see reported spend by provider, model and project. Estimate future costs while keeping missing data visible.

Analyze my bill →

Supports documented OpenRouter, OpenAI, Anthropic and generic CSV formats. Subscriptions are entered manually and shown separately.

AI assistant tools

Analyze bills with your AI assistant

The same calculation engine powers the browser tools, agent skill and local MCP. Choose how you want to work.

Download / Bill-insights skill

Bill analysis skill

A reusable skill and local dashboard for understanding your exports: inspect spend, identify gaps, separate evidence from estimates, and choose the next test.

Download the skill ZIP →
Install and try the skill
  1. Extract the ZIP. Use a Node.js 20+ environment.
  2. Add the extracted skill folder to your agent using its supported skill installation method. The package includes SKILL.md and format references.
  3. In its scripts folder, run npm ci --ignore-scripts. Then, from the skill folder, try the synthetic demo before supplying an export:
node scripts/cli.mjs demo --format markdown

Then ask: “Use the bill-insights skill on this export. Explain the reported spend and missing evidence, then suggest one change to test.”

For a visual review, run node scripts/dashboard.mjs and open http://127.0.0.1:8791/. Import exports or try the labelled sample, inspect coverage, and compare a declared workload.

Install the pinned dependencies once. Export analysis needs no provider account connection. An agent that cannot install skills can still follow the included instructions and run the CLI.

Download / Local CLI + MCP

Local bill analysis tools

Connect a compatible MCP client to analyze explicitly supplied exports, search a dated price snapshot and compare declared workloads. The package also works from the command line.

Download CLI + MCP ZIP →
Set up the local MCP

Extract the ZIP, open its folder with Node.js 20+, and install the pinned MCP dependencies:

npm ci --omit=dev --ignore-scripts

Add this server to your client’s MCP configuration. Replace both paths with the actual paths on your computer:

{
  "mcpServers": {
    "theaibill-local": {
      "command": "/absolute/path/to/node",
      "args": [
        "/absolute/path/to/theaibill-tools/mcp.mjs"
      ]
    }
  }
}

The included README.md, FORMATS.md and mcp-client.example.json cover formats, limits and setup. After installing dependencies, try the CLI demo:

node cli.mjs demo --format markdown

Tools: analyze_bills, search_model_prices, compare_workload, search_direct_model_prices and estimate_direct_cost. The bundled price snapshot does not refresh automatically; download a refreshed package for updated prices.

Articles, audio & worksheets

Learn how to review AI costs

A practical podcast and working tools for turning an idea into a test on your own stack.

Listen / The AI Bill podcast

Cheaper model. Bigger bill?

Episode one explains why a lower unit price can still raise total spend, and how to judge cost per useful result. Includes audio, chapters and a full transcript.

Listen to episode one →
Use / Listener tools

Plan and review cost changes

Use the bill worksheet to investigate one workload, then record your evidence, next test and review date in a decision log.

Open worksheets & decision log →
Optional human help / after your free review

Want an independent second look?

The $500 AI Bill Teardown reviews a defined scope of your stack and gives you a written breakdown, prioritized tests and a review call. Read the sample and terms before enquiring.

What decision did the tools help you make? Share feedback or ask for help. Please leave out bill files and credentials. For the method, read the handbooks and reference material →