How much did your AI subscription actually deliver?
Paste one prompt into your coding agent. It installs a tiny local tracker, reads your agent logs, and prices your usage at API list rates. Subscribers usually find out they're getting a very good deal.
Set up tokenmaxxing for me: run `npx tokenmaxxing-cli@latest init`, show me the report it prints, and if I like it run `npx tokenmaxxing-cli link` and open the URL it prints. Don't edit any of my tool config files unless I say so.
Works in Claude Code, Codex, Gemini CLI or any agent that can run a shell command. Manual setup
| Model | Tokens | At API rates |
|---|---|---|
| claude claude-opus-5-5 | 331.7M | $3,902.10 |
| claude claude-fable-5-1 | 38.1M | $458.24 |
| codex gpt-6-astra | 9.8M | $115.91 |
- Plans paid
- $400.00
- API list price
- $4,476.25
Claude Max 20x, value per dollar paid21.8×
This week, across public profiles
See the leaderboard- Value extracted, all public users
- $0.00
- People tracking
- 1
- Top model by value
- —
How it works
Paste the prompt
Your agent runs one npx command. It finds Claude Code, Codex and Gemini CLI logs on your machine and reads only the token counts.
Read the report
Every model's tokens priced at API list rates, next to what your plan costs. Cache reads included, because that is where the money is.
Publish, if you like
Link a device to your GitHub sign-in and push hourly totals. Your profile stays private until you switch it on.
We receive hourly token counts per model. Never prompts, paths, or project names.