nateherkai

Local dashboard for Claude Code token usage, costs, and session history — reads JSONL transcripts and turns them into charts, heatmaps, and cost analytics.

37
9
100% credibility
Found Apr 21, 2026 at 37 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A local dashboard that analyzes Claude Code session files to display token usage analytics, cost breakdowns, project comparisons, and efficiency tips without sending data anywhere.

How It Works

1
🕵️ Discover the dashboard

While using Claude Code for projects, you learn about a free local tool that tracks your chat costs and usage patterns to help you spend less.

2
📥 Download the files

Grab the project files from the GitHub page and save them in a folder on your computer.

3
🚀 Launch the viewer

Run one simple command, and it scans your Claude Code history, starts a private webpage, and opens it in your browser automatically.

4
📊 See your big picture

Check the overview tab for total sessions, daily charts, project breakdowns, top tools, and estimated costs on your plan.

5
🔍 Dive into details

Switch tabs to review expensive prompts, session histories, skill usage, or get tips like spotting repeated file reads.

💰 Work smarter, spend less

Follow the suggestions to optimize your prompts, cut waste, and confirm you're getting value from your AI subscription.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 37 to 37 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is token-dashboard?

Token-dashboard is a local Python dashboard that scans Claude Code's JSONL session logs from ~/.claude/projects/ and visualizes token usage, costs, and history in interactive charts and tables. Fire up `python3 cli.py dashboard` for a browser-based UI with tabs for overviews, expensive prompts, per-session breakdowns, project comparisons, skill invocations, cost-saving tips, and plan settings—no pip installs, Node.js, or external APIs needed. Everything stays on your machine, with live rescans every 30 seconds and CLI commands for quick stats like today's totals.

Why is it gaining traction?

Unlike cloud token trackers, it's zero-telemetry and stdlib-only, appealing to devs paranoid about prompt privacy, much like airgradient local dashboard or crowdsec local dashboard setups. The hook is actionable insights—spot repeated file reads, bloated tool outputs, low cache hits, or overkill Opus calls—with rule-based tips and Pro/Max plan cost views. Simple one-liner setup beats fiddling with local github copilot alternatives or dashboard token configs in k8s environments.

Who should use this?

Claude Code power users on paid plans optimizing API-equivalent spend, especially those debugging why sessions balloon tokens via tools or subagents. Devs managing multiple projects will love per-project heatmaps and session diffs; prompt engineers tracking cache efficiency across skills fit perfectly. Skip if you're not deep into Claude Code transcripts.

Verdict

Grab it for personal Claude analytics—37 stars and 1.0% credibility reflect early maturity, but stellar README, CLI polish, and zero deps make it reliable now. Solid for solo use; contribute tests to boost it further.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.