Glossary
Usage limits
Usage limits are the caps an AI service puts on how much you can use it: how much you can chat in a set window, file uploads, image generations, or access to the best models. They're the real difference between free and paid plans, and most of them reset on a schedule.
Every plan of every AI service has a meter running somewhere. On a free tier, each company counts in its own way. ChatGPT currently allows unlimited everyday text chats (within anti-abuse safeguards) but puts separate limits on file uploads, image generation and voice. Claude and Gemini use an allowance that resets every five hours, and Gemini adds a weekly cap on top. Paid plans raise the ceiling, but there’s still a ceiling. Features like deep research carry their own caps too, and on the API limits take the form of rate limits and token budgets.
The frustrating part is timing. A plumber drafting quotes with Claude hits the cap mid-afternoon, exactly when the next customer email arrives, and has to wait for the limit to reset. That’s not a malfunction; it’s the business model. Compute is expensive, and limits are how providers keep free and cheap plans sustainable.
Two practical notes. First, limits change often and are sometimes deliberately vague (“more usage”), so treat any number you read as a snapshot. Second, if you keep bumping into the ceiling, that’s data: it tells you which plan you actually need. Our plan picker matches current limits across vendors to how you really work.
Where you’ll meet this
- In-chat banners: “You’ve reached your limit, come back at…”
- Pricing pages, in the fine print under each plan column
- API dashboards, as rate limits and monthly usage meters