In order to ensure high uptime, we impose limits on API requests, even for paid users.
What are the limits?
By default you can make 15 concurrent requests per API key
What is a 'concurrent request'?
This is the number of simultaneous jobs that can be processed at any given time.
> RealTime Request:
This is the number of simultaneous open API connections you can have to our API. Ex: if your limit is 15 and you make 16 API requests at the same time, you will be rate limited on the sixteenth request. However, if one of the 15 jobs finishes early and you make a sixteenth request, that will go through
> Non-RealTime Request:
There is no limit to the number of API calls you can make since you are pushing jobs into a queue with very low latency. However, the speed at which jobs will be processed off that queue will be limited to your rate limit. Ex: if you have 150 jobs and a concurrency of 15, the jobs will be processed in batches of 10.
Can the limit be raised?
Yes, the limits can be raised. There are no strict limits on what it can be raised to since our platform runs on a serverless architecture and can scale efficiently. HOWEVER, we do require you to use the platform and see in reporting that you are near to or exceeding your current limits.
How will I know I'm being rate limited?
We will return an HTTP code 402 with a specific error message related to the exact rate limit that is being exceeded.