Every LiveKit project gets the following for free:
We want to give every project 50GB/60min for free, but beyond that, we need to start charging you. Thus, for your project to use more than your free quotas, you have to enter a payment method and sign up for our Standard plan.
Projects on the Starter plan (free) will stop working after these quotas are exceeded. Projects on paid plans will only be charged for the amount that exceeds these quotas. (For more details, see Pricing.)
LiveKit projects have limitations on the number of concurrent connections of various types in order to ensure the stability of our network and to prevent abuse. Think of them like rate limiting in an HTTP service, but for a continuous service with long-lived connections. Like rate limiting, the primary purpose of these connection limits is to prevent abuse.
You can view the current connection limits on your project at any time in the Cloud Dashboard by navigating to your project’s Settings and clicking the Project tab.
Projects on the Starter plan (free) have the following limits:
- 100 concurrent connections (eg, participants)
- 1 concurrent Egress request at a time
- 1 concurrent Ingress request at a time
If these limits are exceeded, expect new connections of the same type to fail.
Project on a paid plan come with higher concurrent limits. If you think you'll need more concurrent connections for your project, an admin for your project can request an increase for the particular limit in your project settings.
API rate requests
All projects also have a 1000 request per minute rate limit on API requests. Honestly, we don't think any project should exceed this rate limit, but if it’s a problem, please contact support.
Each participant may subscribe to a limited number of tracks. Currently, the subscription limit is:
- Up to 50 video tracks
- Up to 100 audio tracks
We haven't encountered a use case demanding higher limits so far; most end-users usually lack the bandwidth to handle more concurrent tracks.
For video use cases, consider using pagination and selective subscriptions to keep the number of subscriptions within these limits.
For audio use cases, we are developing selective audio muxing in the core SFU, which will selectively transmit active audio data to the subscriber using a manageable number of tracks.