Base vs Max 5x vs Max 20x.

Three plans, same product, differ only in usage cap and meal donation:

- **Ladle ($20/mo)** — the base plan. Fits typical daily use: ~200 messages/day, ~30 files/chat, 200K context. Funds ~10 meals/month. - **Max 5x ($100/mo)** — 5× the usage cap. Fits heavy daily use (thousands of messages, extended coding sessions). Funds ~50 meals/month. - **Max 20x ($200/mo)** — 20× the usage cap. Fits genuinely intensive workflows (constant multi-file coding assistance, large-context research). Funds ~100 meals/month.

**Which one you actually need:** most people are fine on base. If you're not sure — you don't need Max. The signal that you'd benefit is specific: you're hitting the usage cap on base at least 5 times a month and it's actually blocking work (not just being inconvenient). If you're hitting the cap occasionally, wait until it becomes a pattern.

**What "usage" means**: LADLE's cap is expressed as a rolling usage tier tied to tokens processed (both input and output). Long chats, big files, and deep-context Projects consume more; short single-turn queries consume less. The dashboard shows your current usage tier and rolling consumption in Settings → Usage. See "What counts as usage" for the detailed breakdown.

**Why the tier structure at all**: honestly, model inference costs money. A power user on base can consume more inference than the $20 covers on Anthropic's API pricing. The Max tiers are how heavy-usage patterns pay for themselves. If we didn't have them, base would need to be more expensive to average out.

**Switching plans**: instant, prorated. Upgrade takes effect immediately; you're charged the difference for the remaining days. Downgrade takes effect at the next billing cycle. Meal donations adjust proportionally.

**Trial for Max**: none. If you're curious whether Max would help, use base for a full month first and check your Usage dashboard. If you hit the cap 5+ times and it blocked work, Max 5x is defensible. If you didn't, it isn't.