LiveKit review

LiveKit is an open-source framework and developer platform for building, testing, deploying, and monitoring voice, video, and physical AI agents. Its framework gives developers room to customize agent behavior and choose models; LiveKit Cloud provides hosting and scaling for real-time sessions.

The platform covers several pieces of an agent stack: a model gateway for language, speech-to-text, and text-to-speech models; phone numbers and SIP integrations for calls; and session observability. The listed features also include automatic turn detection and interruption handling.

That makes LiveKit a set of building blocks, not a finished customer-support agent you can simply switch on. Teams still need to build the agent and decide which models and integrations fit. The cloud plans start with a free Build tier, while production-oriented usage is priced by plan, session minutes, and some add-ons. For teams comparing providers, the key question is whether that flexibility is worth managing the extra components and costs.

LiveKit puts agent development and real-time communications in one platform. Developers can build voice, video, or physical AI agents, choose models through the gateway, connect calls through telephony, then deploy and observe sessions in LiveKit Cloud. Turn detection and interruption handling are listed features, useful for conversational agents where callers may speak over a response.

The pricing page gives teams a low-cost way to start and clear included session allowances on paid plans:

• Build costs $0 per month and includes 1,000 agent session minutes, five concurrent sessions, one deployment, $2.50 in inference credits, and one free US local number.
• Ship costs $50 per month and includes 5,000 session minutes, 20 concurrent sessions, two deployments, $5 in inference credits, and 5,000 recording minutes.
• Scale costs $500 per month and includes 50,000 session minutes, up to 600 concurrent sessions, four deployments, $50 in inference credits, and 50,000 recording minutes.

Some Reddit users report good call quality or straightforward Cloud setup for their use cases. Others describe self-hosting as requiring command-line knowledge. Those are individual reports, not a guarantee of setup effort or call quality.

The headline plan price is only part of the bill. Ship and Scale charge $0.01 per agent session minute above their included amounts. Recording overages cost $0.005 per minute, and observability events cost $0.00003 each over included amounts. Model credits are limited; additional inference is billed at model prices, which vary by model and plan. Telephony, WebRTC, voice isolation, and turn-handling features can add further usage charges.

The free Build plan is useful for starting, but its 1,000 session minutes, five concurrent sessions, one deployment, and $2.50 in inference credits are firm constraints for a growing or production workload. Ship raises those limits, but still caps concurrency at 20 and includes only 5,000 session minutes. Scale costs $500 monthly; Enterprise pricing and limits are custom. The pricing page lists monthly prices but no annual rates.

There is also integration work to account for. A Reddit user trying to connect LiveKit to an existing application reported difficulty finding suitable integration documentation, while other commenters point to example projects. Self-hosting avoids the Cloud plan price, but users report needing to supply a server and, in some cases, command-line setup knowledge.

LiveKit’s Build plan is free and includes 1,000 agent session minutes, five concurrent sessions, one deployment, $2.50 in inference credits and one free US local number. Ship costs $50 per month: 5,000 session minutes, then $0.01 per minute; 20 concurrent sessions; two deployments; $5 in inference credits; and 5,000 recording minutes. Scale is $500 per month—ten times Ship—with 50,000 session minutes, then $0.01 per minute, up to 600 concurrent sessions, four deployments, $50 in inference credits and 50,000 recording minutes. Enterprise pricing is custom.

The pricing page lists monthly prices but no annual billing or rates. Watch for costs beyond the headline tier: Ship and Scale charge extra for recordings, observability events, phone minutes and some SIP or WebRTC usage. Voice isolation overage is $0.0012 per minute; agent recording overage is $0.005 per minute. Model usage beyond included credits is billed at model prices, which vary. Audio turn detection and adaptive interruption can also incur per-request charges. Price your expected traffic and model mix before moving production agents; the included credits do not make total usage predictable.

Treat a move to LiveKit as an engineering project, not a simple account change. The platform combines an open-source agent framework with LiveKit Cloud, model access, telephony and session observability. Plan to map your agent behavior, model choices, phone or SIP connections and monitoring to the new setup. The published material does not publish a migration tool, data-export process or portability guarantees, so confirm those details before committing. The open-source framework offers a different path from a cloud-only service, but it does not by itself prove that every hosted component or stored record can move cleanly.

Leaving may make sense if your usage regularly exceeds included minutes, model and telephony costs are hard to forecast, or you need a no-code workflow. We were not given Reddit posts or comments to substantiate reasons users leave, so we can’t responsibly attribute those complaints to Reddit. Before switching in either direction, ask LiveKit how to export session records and configuration, and estimate your actual minutes and integrations against the plan caps.