Skip to content
MCP server

Claude Usage

Measures whether a Claude subscription pays off, using local logs and no upload

Pricing
Open source
Platforms
CLI, macOS, self-hosted
Version
1.0.0
Listing updated
today

Similar MCP servers

About Claude Usage

Claude Usage is an open-source MCP server for ai models and llm tooling. It is available on CLI, macOS and self-hosted.

The README says this tool answers a question subscriptions hide: whether you are getting value from a flat-rate Claude plan. It reads local Claude Code logs and status-line samples to estimate how many surviving lines of code your sessions produced, how much spend went to rework or long conversations, and what drives the 5-hour and weekly limits.

It reconstructs edit history from local logs rather than relying on git, so experiments and code rewritten before committing still count. Limit analysis needs a status-line collector that the user adds to Claude Code settings; the project does not edit settings automatically. The author reports measurements from one account over about five weeks and notes that chat usage is estimated by subtraction rather than directly observed.

It suits individual Claude subscribers who want to understand their own usage. The README describes the project as finished and no longer actively developed, and it is limited to Claude.

Tools this server provides (4)

Taken from the project documentation; the running server may differ.

subscription_value
Lines that survived, waste, cost by activity
limit_breakdown
What drives the 5-hour and weekly limits
chat_share
How much of the limit chat ate
current_limits
Limits right now, plus the per-product breakdown

What it does

Surviving lines measure
Counts only code edits still present at the end of a session, folding edits in time order.
Waste cost breakdown
Estimates API-dollar spend lost to rework, failures, rejections and long conversations.
Rate limit drivers
Shows how cache reads, fresh input, output and model choice affect the 5-hour and weekly limits.
Chat usage estimate
Estimates how much limit claude.ai chat and the mobile app consumed, by subtracting logged usage from the gauge.
Local-only operation
Reads local logs and status-line samples, with no upload; the usage command sends credentials only to Anthropic's endpoint.

Good for

  • Checking whether a Claude subscription is worth its price for personal coding work
  • Finding out why the 5-hour or weekly Claude limit drains quickly
  • Estimating how much limit goes to claude.ai chat versus Claude Code
  • Reviewing long conversations that inflate token costs

Pricing

Open source

Free and MIT licensed; runs locally.

Claude Usage FAQ

Does it upload my data?

The README states there is no upload and that logs are read locally. The live usage command sends stored credentials only to api.anthropic.com.

Does it need a Git repository?

No. The README says it reconstructs edit history from local logs and does not use Git integration.

What platforms does it run on?

It requires Python 3.9+ with the standard library only. The README says macOS is required for the live-limit tool.

Is it still maintained?

The README marks the project as finished in October 2026, no longer developed, with issues possibly unanswered.

Is Claude Usage free?

Yes. Claude Usage is open source, so you can use and self-host it for free.

What are the best alternatives to Claude Usage?

The closest alternatives to Claude Usage are MCP Claude Usage Guard, context-doctor, Kestrel, Claude API Pricing and Widal.Code.

Where this comes from

  • Description: written from the project's README and its MCP Registry entry.

How AIPoweredBot checks servers