Viktor.comAn AI coworker that actually does the work
Promoted
TrackNotch
LLM usage tracking that lives in your Mac's notch
99 followers
LLM usage tracking that lives in your Mac's notch
99 followers
TrackNotch is a native macOS app that shows Claude, OpenAI, Cursor, and Codex usage right in your notch. Context arc, budget tracking, rate limits and all local. No proxies, no telemetry, nothing leaves your machine. Free and open-source.
TrackNotch
the budget tracking piece is interesting for teams but right now it sounds personal use only. is there any plan for shared visibility, like if a small team wanted to track aggregate spend across multiple machines without everyone having to check their own notch individually. or is keeping it single machine intentional because the moment you add sync you have to make privacy tradeoffs you've explicitly avoided
TrackNotch
@ansari_adin I’m intentionally keeping TrackNotch single‑machine and local‑only right now. The moment I add sync/shared dashboards, I’d need a backend and would have to make privacy tradeoffs I’m trying to avoid. For org‑wide API spend, I imagine teams using their existing billing/observability tools, and TrackNotch staying focused on power‑user visibility on each Mac.
The local-only processing is the right call for an API key-adjacent tool. Anything that proxies or phones home with usage data is a non-starter for a lot of developers. One thing I'm curious about: how does TrackNotch get read access to the token counts? Is it reading from the local provider SDKs' log files, intercepting at the network layer, or pulling from the official usage APIs on a poll interval? The answer affects both accuracy and whether it works for team accounts vs. individual API keys.
TrackNotch
@binu_george For the local tools (Claude Code, Cursor, Codex), TrackNotch just reads the usage and state they already write to disk on your machine. No network interception, no proxying; it is piggybacking on the same local data those apps themselves use.
For raw API spend, it talks to the official Anthropic and OpenAI usage endpoints on a short poll interval using whatever credentials you give it. That is what makes the split you mentioned: editor integrations are inherently per machine and individual, whereas cost and budget tracking is org level and only works when you supply an admin or org key that can see aggregate spend.
So accuracy is bounded by what the tools and providers themselves record, and the team versus individual story maps cleanly onto “local editor state” versus “org billing APIs".
That is really good.
Does it track calude code/antigravity costs or usage
TrackNotch
@jay_gangwar Thank you and yes it does track both of them.
@manoj_achari That is wonderful!
@manoj_achari what is the pricing ?
TrackNotch
@jay_gangwar it is free, you could just head to the git and click on releases download the zip and you are good to go.
Wow, interesting, will check this out, I was looking for something like this. Does it support gemini/codex too ?
TrackNotch
@vinitvr Yes not gemini but antigravity and codex.
@manoj_achari , yeah antigravity is fine. gemini cli is more or less retired now.
TrackNotch
@vinitvr True that.
Really nice github project and super nice your giving out. Gave the repo a star :) . Its actually something i would use a lot i live on the edge of my claude limit lol and am tired of refreshing the site every 30 minutes to see how close i am. Will give a try really cool stuff!
TrackNotch
@andrewb23 Yeah that was that was my anxiety as well, thank you for the star ^_^
Love that it lives directly in the Mac notch.
TrackNotch
@nithin_raju1 Thank you! I am glad you found it useful!