@binu_george For local tools (cloud code, cursors, codecs), Tracknaut only reads the usage and reports that they are already writing to disk on your machine. No network blocking, no proxy; It’s based on the same local data that these apps themselves use.
For raw API spend, it talks to the official Anthropic and OpenAI usage endpoints over a short polling interval using any credentials you provide. That’s the split you mentioned: Editor integrations are naturally per machine and person, while cost and budget tracking are at the organization level and only work if you provide an administrator or organization key that can see total spending.
So accuracy depends on what the tools and providers themselves record, and the team vs. individual story maps clearly onto “local editor status” vs. “organization billing API”.
<a href