Free token usage analyzer

What's actually driving your Claude Code session's token usage?

Paste a few lines from a session transcript and see the breakdown by cause - conversation text, tool output by type, and duplicate reads - not just a quota number.

Paste lines from a session transcript (.jsonl)0 chars

Paste a transcript excerpt and click Analyze to see the token breakdown by cause - conversation text, tool output by type, and any duplicate file reads.

This only covers what you pasted, nothing leaves your browser. Run the free scan for every session in your real history, computed the same way, automatically.

npx usagecutRun a free scan
Recommended next step

A clean setup is just the beginning.

UsageCut trims that for free. Want Claude Code to be genuinely better, not just leaner? ClockedCode is the curated, conflict-free setup - installed in one paste.

Upgrade your Claude Code
  • 10x your Claude Code with hand-picked tools
  • Specialist agents on your team
  • Stop spending hours researching plugins
Why a breakdown, not a total

The same categorization UsageCut's scan runs on your real history

Claude Code's own /usage and /context commands are quota views - they tell you the total is big, not why, as how many tokens Claude Code is actually using covers. This tool runs the same per-cause split UsageCut's local scan computes on a full history - conversation text versus tool output, tool output attributed back to the call that produced it (a Read, a Bash command, a subagent), and any file read twice in the same excerpt - against just the lines you paste in.

Paste the built-in sample above to see it running on a real 4-line exchange: two duplicate Read calls against the same file, a Bash test run, and one plain reply. For the CLAUDE.md side of the same problem, see the CLAUDE.md token counter - or run the free scan below for every session in your real history, computed automatically.

Free tools

More free tools for Claude Code

Small, focused, no-signup tools from the maker of ClockedCode. Free to use, forever.

UsageCutScan your real Claude Code setup and see exactly where your tokens are wasted, then cut it losslessly.Claude Code Cost CalculatorEstimate your monthly Claude Code cost from sessions per day and workload, and see whether Pro, Max, or API pricing wins.CLAUDE.md Token CounterPaste your CLAUDE.md or AGENTS.md to count its tokens and see which sections look conditional enough to move out.MCP Server Token Cost CheckerPaste your mcpServers config to see the estimated token cost per server, worst first, with tool-search deferred loading on or off.CLAUDE.md GeneratorAnswer a few questions about your stack and get a CLAUDE.md scoped to just that, with a live token cost per section.Context Window CalculatorSee what percent of your context window is used, your headroom before compaction, and the cache-read cost of carrying it.Usage Limit CalculatorProject your Claude Code usage-limit headroom across Pro, Max 5x, and Max 20x from your own baseline and Anthropic's confirmed usage multiplier.Subagent Cost CalculatorSee whether spawning N Claude Code subagents actually saves tokens versus doing the same work in one session, using the measured 216-token subagent overhead.Best MCP Servers for Claude CodeBrowse popular Claude Code MCP servers by category and see each one's estimated token cost per session, sorted worst to best.Hook Token Cost CalculatorPaste your hooks config to see the estimated token cost per hook, worst first, classified by how often each event actually fires.Tool Search Token Savings CalculatorPaste your mcpServers config to see how many tokens Claude Code's tool search saves by deferring schemas, and whether a custom auto:N threshold would load them upfront instead.Opus vs Sonnet Cost CalculatorSee the dollar gap between running Claude Code on Opus vs Sonnet, plus the honest (not naive) savings from pinning just the Sonnet-safe turns.Claude Code Model PickerAnswer 5 questions about your current task to get a straight verdict: stay on Opus, hand it to a Sonnet subagent, or use opusplan.Which Claude Plan Should You UseAnswer a few questions about your usage and team size to get a straight Pro, Max, Team, or Enterprise recommendation with real current pricing.Compaction Cost CalculatorSee how many times a Claude Code session will hit auto-compact and what each compaction actually costs in tokens, cache-warm or cold.

More free Claude Code tools are on the way - built one at a time by a solo dev who ships daily.

Built for developers using Claude Code every day - by a solo dev who ships daily and uses this on his own setup.

Open-source scanner, runs locally, your code never leaves your machine.

Claude Code tips, every Sunday

One short email a week: the token-saving tricks, setup tweaks, and tools worth your time. Free, unsubscribe anytime.

Token usage analyzer FAQ

About this breakdown

Where do I get a transcript to paste?

Claude Code writes one JSONL file per session to ~/.claude/projects/<project>/<session-id>.jsonl (or the equivalent under %USERPROFILE%\.claude\projects on Windows). Open one in a text editor and copy a few lines, or run something like `tail -100 ~/.claude/projects/*/*.jsonl` in a terminal and paste the output.

Is this an exact token count, like the real bill?

No - it's the same chars/4 heuristic estimate UsageCut's own local scan uses, not Anthropic's actual tokenizer or billed usage. It's also scoped to only what you paste, not your full history. For your real, per-project numbers across every session, run the free scan.

Why does it flag duplicate reads?

Because re-reading a file that's already in context re-bills that same content on every later turn instead of reusing what's already there - it's one of the few token-waste signals cheap enough to compute from a short excerpt. See how to reduce Claude Code's context window usage for the rest of the levers.

Does this analyze my whole history, or just what I paste?

Just what you paste, and it never leaves your browser. For your whole history, run `npx usagecut` locally - it reads every session file directly, computes this same breakdown automatically, and needs no copy-pasting.

FAQ

Questions, answered

Is UsageCut really free?

Yes. The scan and every fix are free - no card, no subscription, no signup. You only give an email if you want the optimization plan and one-command undo sent to your inbox.

Is my code safe? Does UsageCut upload anything?

The scan runs entirely on your machine. It never uploads your code, never reads your API key, and never sends your prompts or conversations anywhere. Only anonymous summary counts leave, and only if you choose to share them.

How does it reduce Claude Code token usage?

It reads your local setup and session history to find waste - idle MCP servers loaded into every session, an oversized CLAUDE.md re-sent on every request, duplicate file reads, and bloated tool output - then trims it losslessly, so each session carries less context and you hit your limit far less often.

What is ClockedCode?

ClockedCode is the curated, conflict-free Claude Code setup from the same maker. UsageCut makes your current setup leaner; ClockedCode makes it genuinely better - a set of hand-picked tools, specialist agents, and a tuned CLAUDE.md, installed in one paste.

Who is ClockedCode for?

Developers who use Claude Code daily and want a setup that is powerful out of the box instead of spending hours researching and wiring up plugins, agents, and MCP servers themselves.

What's included in ClockedCode?

A vetted, conflict-free bundle of hand-picked tools, specialist agents, and a tuned CLAUDE.md - everything installed together in one paste, with no setup archaeology.

Is ClockedCode a one-time payment?

Yes. ClockedCode is a one-time purchase with an instant download - no subscription and no recurring fees.

Does ClockedCode work with my existing setup?

Yes. It installs alongside what you already have and is fully reversible. Run UsageCut first to clean your current setup, then add ClockedCode to level it up.