Claude Code: /usage won't show your remaining quota outside a session — how to calculate the limit yourself

⚡ The gist in 5 seconds - In an article on Habr, the author describes a method for forecasting quota consumption in Claude Code, Codex, and Kimi Code until the window resets. - Claude Code has two separate limits: a five-hour session window and a weekly window — you need to track both. - The built-in indicator doesn't solve the problem: /usage shows the remaining quota only inside an open session and only for a single account. ### 🔍 What was found The Habr article ("Will the quota last until the reset? How I forecast Claude Code, Codex, and Kimi Code usage") tackles a practical problem faced by code-agent subscribers: the /usage command in Claude Code reports the remaining quota only within an already-open session and only for one account. The author notes that the subscription has two independent limits — a five-hour window and a weekly one — and it's precisely these that have to be compared to figure out whether your quota will last until the reset. The forecasting method is built around manually tracking consumption: instead of a one-off quota check, you need to monitor usage dynamics in both windows and extrapolate them to the reset point. The article applies the same approach to Codex and Kimi Code — that is, to multiple services at once, which makes the built-in tools of a single account insufficient. ### 💡 Why it matters Code-agent subscription limits are a real bottleneck: if you hit the quota mid-task, you lose access to the model until the window resets, and the non-obvious difference between the five-hour and weekly limits only complicates planning. Until vendors provide a unified indicator across multiple accounts and services, users have to calculate forecasts manually — and the article describes exactly such an approach. This is a user-made methodology, not an official Anthropic feature, so don't expect it in a stable release.