fien-atone

See where your AI coding money goes. Self-hosted dashboard for Claude Code, Cowork, and Codex CLI — all data stays on your disk.

18
1
100% credibility
Found Apr 24, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

A self-hosted dashboard that parses local session files from AI coding tools like Claude Code, Claude Desktop, Cowork, and Codex CLI to visualize costs, usage patterns, and activity breakdowns.

How It Works

1
🔍 Discover Third Eye

You hear about a simple tool that reveals exactly how much your AI coding assistants like Claude are costing you, with charts showing spend by project, model, and time.

2
📥 Get it set up

Choose an easy way to install—use Docker for simplicity, Node if preferred, or even ask your AI to handle it automatically.

3
🚀 Launch on your computer

Start the dashboard with one command, and it connects to your local AI session files without sending data anywhere.

4
📊 See your spending at a glance

Open your browser to a beautiful dashboard filled with charts breaking down costs by model, project, activity, and when you work.

5
🔍 Dive into projects

Click any project to explore details like top files edited, tools used, git branches, and activity heatmaps.

6
⚙️ Customize your view

Drag, resize, and add widgets to tailor the dashboard to what matters most to you, saving layouts locally.

💰 Take control of your AI budget

Now you know exactly where your money goes, spot wasteful patterns, and make smarter choices for your coding workflow.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 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 third-eye?

Third Eye is a self-hosted TypeScript dashboard that scans your local session files from Claude Code, Cowork, and Codex CLI to reveal exactly where your AI coding money goes. It delivers charts breaking down costs by model (Opus vs. Sonnet), project, activity type, git branch, and time—down to per-call granularity, including ephemeral sessions provider dashboards miss. Fire it up via Docker or Node, and all data stays on your disk in SQLite, with no cloud or signup needed.

Why is it gaining traction?

It stands out by surfacing hidden spend patterns—like prompt caching savings or debugging vs. new code costs—that vendor portals gloss over, all in customizable, drag-resizable widgets with heatmaps and project drill-downs. Multi-language support, themes, and shareable URLs make it instantly usable, while Docker auto-ingests every 15 minutes for fresh views without manual hassle. Developers love the "aha" of seeing GitHub Copilot usage equivalents in their Claude flows, locally.

Who should use this?

Claude Code or Codex CLI heavy users tracking monthly AI bills across projects. Ideal for solo devs or teams optimizing token burn on specific repos, or anyone curious about work patterns like peak hours via activity heatmaps. Skip if you're not deep into these tools.

Verdict

Worth spinning up for Claude loyalists—functional Docker setup and detailed docs punch above its 18 stars and 1.0% credibility. Early maturity means watch for edge cases, but it's a sharp, private alternative to vague billing pages.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.