Starter
For getting a real product in front of real users.
- 1.5M participant-minutes
- 3 TB of bandwidth
- 5,000 concurrent participants
15 blocks · overage never exceeds 20% above your plan rate
Talk to us about StarterParticipant-minutes and bandwidth, on infrastructure we run ourselves. No per-seat licence, no negotiation to see a number.
For getting a real product in front of real users.
15 blocks · overage never exceeds 20% above your plan rate
Talk to us about StarterFor an app that has found its audience.
30 blocks · overage never exceeds 20% above your plan rate
Talk to us about GrowthA cluster of your own, shared with no one else.
80 blocks · overage never exceeds 20% above your plan rate
Talk to us about ScaleFor platforms where streaming is the product.
200 blocks · overage never exceeds 20% above your plan rate
Talk to us about PeakTwo meters, and both of them are yours to check. Every plan includes an allowance of each; you are only charged beyond what you have already bought, and the two allowances are independent — running out of one does not consume the other.
| What we measure | How |
|---|---|
| Participant-minutes | Every participant, every minute they are connected |
| Bandwidth | Media delivered from our servers to your users |
Going over costs at most 20% more per unit than your plan rate — the same on every tier. Most providers price overage to punish you into upgrading. We would rather you upgraded because it was cheaper, which it always is.
Minutes used, bandwidth used, what was included, and what each unit cost — line by line, so you can check any figure against your own logs.
Rooms, participants, recordings of your own signalling, API calls — none of it is billed. There are exactly two meters, and they are the two above.
Every plan on this page is a starting point, not a ceiling. If your traffic sits between two tiers, runs in one region, spikes on a schedule, or is simply larger than the top card, tell us the shape of it and we will price it properly.
You will be talking to the people who run the clusters, not to a sales desk reading from a script.