Early Access — Leaderboard is Live

You're burning tokens.
Might as well keep score.

Token Hoggers reads your local AI tool logs — Claude Code, Cursor, Gemini, Ollama — and turns raw usage into XP, tiers, streaks, and a global leaderboard. Think GitHub contribution graph meets D&D character sheet.

Your prompts stay on your machine. We only count.

See the Leaderboard Source Code
$ pip install tokenhoggers && tokenhoggers status
30 seconds to your first stats

Install. Check your score. Flex on the board.

01

pip install tokenhoggers

One command. Parses local log files from Claude Code, Cursor, Gemini CLI, and Ollama. macOS, Linux, WSL. Zero config.

02

tokenhoggers status

Instantly see your token count, message volume, and usage streaks — no account, no cloud, no network call. Just your machine.

03

tokenhoggers sync

Sign up, sync, and watch your RPG profile come alive — XP, levels, tier badges, and a rank on the global leaderboard. Now it's a competition.

The Board

Who's hogging the most tokens?

Rankings

All Time Daily Weekly
# User Tier Tokens Level Streak
1 sudo_vibes Gold 2,847,120 18 34d
2 context_window_maxer Gold 1,923,500 15 21d
3 yolo_deployer Silver 891,200 11 14d
4 async_await_what Silver 654,800 9 7d
5 cursor_tab_addict Bronze 312,400 6 5d
6 llm_whisperer Bronze 148,900 4 2d
7 you? Bronze — — —
Claim your spot
Tool Agnostic

We don't care which AI you use. We count all of them.

Claude Code
Cursor
Gemini CLI
Ollama
ChatGPT (soon)
Copilot (soon)
Your Character Sheet

Raw token count is boring. This is an RPG.

Your usage maps to the AI Maturity Index (AMI) — four dimensions that measure how you use AI, not just how much. Each dimension is a stat on your character sheet. Grind one, or balance all four.

B
Breadth
Multi-tool user or one-trick pony?
E
Efficiency
Output per token. Do more with less.
A
Active
Daily streaks. Consistency compounds.
S
Sophistication
Tool calls, agents, multi-turn — power user stuff.