The tools run on servers shared by everybody, so use is counted in credits. These are the limits for use without an account, and for an account on the free plan. Pricing has what Personal gets, and what each level can use has the tools and models.
| limit | |
|---|---|
| storage | 8 MB in the browser, or 512 MB with an account; 2,000 files and folders, and 95 MB for one file in an account |
| one call | 95 MB sent at once |
| a run | 2,000 files; a folder stands for every file in it, subfolders too |
| a zip archive | 128 MB and 2,000 files unpacked, or packed, at once, and never more than your room |
| a document read as sheet music | 50 pages at once |
| credits | 5 a week by address, or 30 with an account, or 200 a day on Personal: a credit is a page read, or a piece engraved, aligned or cleaned up |
| calls at once | 2 for one person |
| a flow sent to the API | 4 uses of a tool; longer batches run from the workspace, or one call each |
| task history | the last 200 runs, kept for 30 days |
| uploads and results through the API | kept for a quarter of an hour |
Credits are counted as each call starts. A call that needs more than is left is refused before it runs and costs nothing, and the workspace checks a whole run against what is left before it starts one.
When the servers are busy, a task waits, says so on the tasks page, and tries again by itself. Nothing needs doing. There is also a limit for each day that everyone on the free plan shares. If it has all been used, the message says so and the tools work again the next day. Personal has its own, apart from it.
Reading a whole collection? For institutions says how the limits add up.