[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"readme:tare":3},"\u003Ch1>tare\u003C\u002Fh1>\n\u003Cp>\u003Cstrong>Ask Claude Code where your usage went.\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cp>You hit a usage limit and don't know why. Your quota drains faster than it\nused to. You suspect something is eating tokens in the background. The\nrecords that answer all of this are already on your computer — Claude Code\nkeeps a log of every request it makes. \u003Ccode>tare\u003C\u002Fcode> teaches Claude Code to read\nits own logs, so you can just ask.\u003C\u002Fp>\n\u003Cp>No dashboards, no commands to learn. Install once, then ask in plain\nEnglish.\u003C\u002Fp>\n\u003Cp>\u003Cem>Tare: the weight of the container, subtracted to find what's inside. Most\nof what a session costs is the container — context re-sent again and again —\nand that's exactly what these tools subtract.\u003C\u002Fem>\u003C\u002Fp>\n\u003Cp>Read full story - \u003Ca href=\"https:\u002F\u002Fwww.kelviq.com\u002Fblog\u002Fclaude-code-usage-limits-where-tokens-go\u002F\" rel=\"nofollow ugc noopener\">https:\u002F\u002Fwww.kelviq.com\u002Fblog\u002Fclaude-code-usage-limits-where-tokens-go\u002F\u003C\u002Fa>\u003C\u002Fp>\n\u003Ch2>Install\u003C\u002Fh2>\n\u003Cp>One command, in any terminal:\u003C\u002Fp>\n\u003Cpre>\u003Ccode class=\"language-bash\">npx skills add kelviq\u002Ftare -g -y --copy --agent claude-code\n\u003C\u002Fcode>\u003C\u002Fpre>\n\u003Cp>That's it. Nothing else to set up — no accounts, no packages, no\nconfiguration. Start a \u003Cstrong>new\u003C\u002Fstrong> Claude Code session and it's live — type\n\u003Ccode>\u002F\u003C\u002Fcode> and check that \u003Ccode>tare\u003C\u002Fcode> appears. If it doesn't, see the\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fkelviq\u002Ftare\u002Fblob\u002FHEAD\u002FINSTALL.md#troubleshooting\" rel=\"nofollow ugc noopener\">troubleshooting note\u003C\u002Fa> — on machines where\nClaude Code has never installed a skill before, the installer can miss\nit. (Other ways to install — by hand, for a whole team, or as a Claude\nCode plugin — are in \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fkelviq\u002Ftare\u002Fblob\u002FHEAD\u002FINSTALL.md\" rel=\"nofollow ugc noopener\">INSTALL.md\u003C\u002Fa>.)\u003C\u002Fp>\n\u003Ch2>Then just ask\u003C\u002Fh2>\n\u003Cp>Open Claude Code and ask your question the way you'd ask a person. These all\nwork — the words don't have to match, complaining about your limits is\nenough:\u003C\u002Fp>\n\u003Cp>\u003Cstrong>When you hit a limit\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Why did I hit my usage limit yesterday?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>I got locked out ten minutes into my evening session — how is that\npossible?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Did I hit the 5-hour limit or the weekly cap?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Why am I burning through my quota so much faster this week?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cp>\u003Cstrong>Where your tokens go\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Where did my tokens actually go this week?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Which of my projects is eating my quota?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Which model is costing me the most?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>What's the most expensive file Claude keeps re-reading?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>How much did that giant session yesterday actually cost me?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Are my MCP servers adding a lot to my context?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>How much overhead do subagents and skills add?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cp>\u003Cstrong>Before you start something big\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cblockquote>\n\u003Cp>\u003Cem>How full is my 5-hour window right now?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Is it safe to start a big refactor now, or should I wait for my window to\nclear?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cp>\u003Cstrong>Checking for things you forgot\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Is something running Claude Code in the background?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Was Claude Code active while I was asleep?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>I set up an automation last month — what is it costing me?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cp>\u003Cstrong>Tuning your setup\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cblockquote>\n\u003Cp>\u003Cem>I started using \u002Fclear between tasks — did it actually help? Compare this\nweek to last.\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Did the latest Claude Code update change my usage?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>What would my usage cost if I were paying for the API directly?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>What one change would save me the most?\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cp>\u003Cstrong>Reports and sharing\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Make me a usage report I can open in my browser.\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Give me a summary I can post publicly — with nothing private in it.\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cblockquote>\n\u003Cp>\u003Cem>Export my usage to a spreadsheet.\u003C\u002Fem>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cp>Prefer typing commands? \u003Ccode>\u002Ftare\u003C\u002Fcode> runs the full diagnosis, and takes variants\nfor the common asks — asking in plain words always works too:\u003C\u002Fp>\n\u003Ctable>\n\u003Cthead>\n\u003Ctr>\n\u003Cth>\u003C\u002Fth>\n\u003Cth>\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\u003Ctr>\n\u003Ctd>\u003Ccode>\u002Ftare\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>full diagnosis — where tokens went and why\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Ccode>\u002Ftare usage\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>at-a-glance panel — like \u003Ccode>\u002Fusage\u003C\u002Fcode>, with attribution\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Ccode>\u002Ftare window\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>how full is the 5-hour window — safe to start?\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Ccode>\u002Ftare report [days]\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>build the HTML report and open it\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Ccode>\u002Ftare tools [days]\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>what is filling my context\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Ccode>\u002Ftare week\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>compare this week with last\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Ccode>\u002Ftare share [days]\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>redacted summary safe to post publicly\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>\u003Ccode>\u002Ftare why did I hit the limit yesterday\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>any question works as the argument\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\u003C\u002Ftable>\n\u003Cimg width=\"2082\" height=\"5376\" alt=\"image\" src=\"https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F24809e7c-065d-4ec1-99d1-d7ae44dd4bb3\" \u002F>\u003Ch2>What you get back\u003C\u002Fh2>\n\u003Cp>Not a wall of numbers — a cause. The answer to \"why did I hit my limit\nyesterday?\" looks like this:\u003C\u002Fp>\n\u003Cblockquote>\n\u003Cp>**99% of yesterday's usage came from a tool you're running,\u003C\u002Fp>\n\u003C\u002Fblockquote>\n",1788219551206]