How usage and limits work
Every Dropstone plan comes with a weekly usage allowance shared across chat, the CLI, and the SDK. How it is measured, when it resets, what happens when you reach it, and how to make it go further.
Your plan includes a weekly allowance of usage. Everything you do with Dropstone, on every surface, draws from the same allowance.
How usage is measured
Usage is measured in credits. Each request costs credits based on how much work it involves: how long the input is, how long the reply is, and which model handled it. A short question to Dropstone Fast costs very little. A long refactor on Dropstone Heavy costs a lot more.
You do not need to track credits by hand. Your dashboard shows how much of the week's allowance you have used, and Dropstone Chat and the CLI show it in the interface.
The weekly window
Your allowance covers a rolling seven-day window. The window opens with your first request after the previous one ends, and runs for seven days from then. When it ends, your allowance resets in full.
This means the reset is tied to when you use it, not to a fixed day of the week.
One allowance, every surface
Work in chat, the CLI, the VS Code extension, and the SDK all count against the same allowance, because they are the same account. There is no separate CLI quota. If you use the CLI heavily on a Monday, that is allowance not available in chat on Tuesday.
What happens at the limit
When you reach the allowance, Dropstone tells you, shows when the window resets, and offers what you can do:
- Wait for the reset.
- Upgrade to a plan with a larger allowance, which takes effect immediately.
- Top up with usage credits, on Pro and above, for weeks when you need more than the plan includes.
Nothing is lost. Your conversations, memory, and projects are all there when you continue.
Making it go further
Use Fast by default. It handles most work and costs the least per request. Reach for Pro or Heavy when you need them. See Choose a model.
Keep conversations focused. Every message in a conversation is part of the input to the next reply. A very long conversation costs more per message than a fresh one. Starting a new conversation for a new topic is cheaper, and memory carries the context across anyway.
Let memory do the repeating. Once Dropstone has learned a preference, you do not need to restate it. See How Dropstone remembers you.
Attach what you need. A large document is read in full each time it is part of a conversation. Attach the part you need, or put it in a project and ask about the relevant section.
Checking your usage
Open Usage and analytics in the dashboard for a breakdown by day, by surface, and by model, and for when the current window resets.