TokenTrack lives in your menubar and shows, in real time, the API-equivalent value of every token you burn — across Claude Code, OpenAI, and DeepSeek. Know if you're getting $200 or $2,000 out of that $200 plan. 100% local.
v0.6.0 · macOS 14+ · notarized · free tier + 14-day Pro trial · $199 lifetime for the first 100
ccusage is great — when you remember to run it. TokenTrack is the always-on, multi-provider, alert-driven layer on top of the same accurate numbers.
Today / week / month / 30-day API-equivalent spend, broken down by model and project, always in your menubar.
Set a monthly cap. Get nudged at 70%, 90%, and 100% — before you hit a surprise, not after.
Claude Code, OpenClaw, OpenAI, and DeepSeek aggregated into one number. Add any model's rate yourself.
Your usage data never leaves your Mac. No account required to track. Reads the logs already on your disk.
Handles message dedup, 1-hour vs 5-minute cache pricing, and new models — reconciled against ccusage.
Optionally syncs the model price table from the LiteLLM registry, so new models are priced the day they ship.
Free forever: track any one AI source (Claude Code, OpenAI, or DeepSeek), with accurate per-model pricing — no account, no card. Every download starts with a 14-day trial of Pro (all features). Upgrade for all providers at once, real-time remaining-budget + alerts, and auto price sync.
TokenTrack is in private beta. Drop your email and we'll send you the signed build plus a founder lifetime offer.