Claude Code hitting Max limits too fast — is it you or the platform? (shareable check)

#16157 "Instantly hitting usage limits with Max" has 720+ reactions. Before you conclude it's a bug, the first thing most people miss: your Pro/Max quota is shared across Claude and Claude Code. This check sorts how much of your fast burn is self-explainable from a possible anomaly worth reporting. Nothing leaves your browser.

Your plan

Other Claude surfaces you use heavily on the same plan (these share the quota)

Anthropic's docs: usage limits are "shared across Claude and Claude Code — all activity in both tools counts against the same usage limits." A heavy morning in the claude.ai app eats directly into your Claude Code window.

Default model for routine work

Typical session shape

This card sorts cause from anomaly — it deliberately doesn't read your quota, because Claude Code exposes no real-time quota number (#13585). Measure your real consumption for free with ccusage, read the full why-am-I-hitting-limits explainer, and cut the burn for free with the 13 cost-spike patterns reference (a concrete defense for each). To go further, the Token Book — English edition (¥2,500+, name your price) adds the per-run token-reduction playbook worked end to end — cache hygiene, model routing, and context trimming.

Hitting Max limits too fast — you or the platform?