LiveKit Cloud
Default when the agent is a realtime session (voice or video) you want to own. Vapi for turnkey telephony…
Default when the agent is a realtime session (voice or video) you want to own. Vapi for turnkey telephony agents; Cartesia/Deepgram as leaves inside LiveKit Inference. Strengths: Best OSS+Cloud realtime media substrate for voice/video agents; Transparent per-min inference catalog; 99.99% network + region pinning / HIPAA on Scale.
Lock-in risk: higher = more lock-in (worse). Overall already inverts this dimension.
Default when the agent is a realtime session (voice or video) you want to own. Vapi for turnkey telephony agents; Cartesia/Deepgram as leaves inside LiveKit Inference.
Strengths
- · Best OSS+Cloud realtime media substrate for voice/video agents
- · Transparent per-min inference catalog
- · 99.99% network + region pinning / HIPAA on Scale
Weaknesses
- · You still compose STT/LLM/TTS (more assembly than Vapi)
- · Plan limits on concurrent sessions can surprise
- · Self-host ops burden if you skip Cloud
Pricing notes
As of 2026-09-14 (livekit.com/pricing): Build $0 (1k agent minutes, $2.50 inference credits, 1 free US number); Ship from $50/mo (5k agent min then $0.01/min, $5 credits); Scale from $500/mo (50k min, $50 credits, HIPAA/region pin); Enterprise custom. Example all-in ~$0.067/min (LLM+STT+TTS+obs+$0.01 session). WebRTC $0.0004–$0.0005/min overage. OSS Agents framework + media server self-host free. Inference catalog includes Cartesia/Deepgram/AssemblyAI/Rime (some TTS SKUs free on LiveKit).
Sourced pricing
Contract status unavailable · no_sourced_price. Single normalized $/1M: null. Machine list: /api/pricing?provider=livekit
No sourced SKU row. Notes above stay prose.
Community signals
Insight claims only — never mutate official scores. Outside reports: 0; seed examples: 0.
No insights yet. Submit a production claim →
All insights for this provider →