Vendor intelligence
LiveKit
Realtime agent infrastructure for teams that own the application and model stack.
Read the signals. Keep the scope.
A place in your stack.
Best for
- Engineering teams building custom phone and web voice workflows.
- Teams needing explicit control over speech, model and carrier choices.
Watch-outs
- Agent-session pricing excludes a complete inference and telephony stack.
- Budget for engineering, operations and production configuration; free-tier allowances do not establish production capacity.
Product specification
What the evidence says.
Unknown means we have not established it. It does not mean No. Compliance claims require plan and contract review.
White label
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)Client tenancy
Project isolation and Enterprise consolidated workspace billing do not establish a reseller subaccount hierarchy or hosted white-label custom-domain feature.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)Source ↗ (opens in new tab)Client billing
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)ISO 27001
Not established by the evidence reviewed. This does not mean unsupported.
No verified sourceBAA availability
BAA available on Scale/Enterprise; buyer must execute and configure covered service.
Source ↗ (opens in new tab)SIP
Inbound/outbound SIP trunks and dispatch setup required.
Source ↗ (opens in new tab)Warm transfer
Python WarmTransferTask is beta; outbound trunk required; custom workflow can return to caller when manager unavailable.
Source ↗ (opens in new tab)Cold transfer
SIP REFER closes LiveKit session; provider must allow transfers.
Source ↗ (opens in new tab)Batch calling
Not established by the evidence reviewed. This does not mean unsupported.
No verified sourceMCP tools
Agent tools can use MCP servers; not evidence of every named integration.
Source ↗ (opens in new tab)Webhooks
Server APIs and webhook room/participant events are documented; HTTP action tools execute configured application integrations.
Source ↗ (opens in new tab)Calendar
Mistral Connectors plugin can access Google Calendar with a configured connector_id; not a native LiveKit calendar account integration.
Source ↗ (opens in new tab)CRM
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)Bring your carrier
Inbound/outbound SIP trunks and dispatch setup required.
Source ↗ (opens in new tab)Bring your model
Model plugins or managed Inference selection; component compatibility needs checking.
Source ↗ (opens in new tab)Bring your voice
Model plugins or managed Inference selection; component compatibility needs checking.
Source ↗ (opens in new tab)HIPAA / BAA
BAA available on Scale/Enterprise; buyer must execute and configure covered service.
Source ↗ (opens in new tab)SOC 2
Supplier states independent audit; actual audit artifact gated under NDA, not reviewed.
Source ↗ (opens in new tab)GDPR
Supplier alignment statement.
Source ↗ (opens in new tab)Languages
Selected multilingual STT models auto-detect and code-switch; selected TTS must also support caller language.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)Voices
Not established by the evidence reviewed. This does not mean unsupported.
No verified sourceRealtime audio
No latency benchmark performed.
Source ↗ (opens in new tab)Voice cloning
Not established by the evidence reviewed. This does not mean unsupported.
No verified sourceArchitecture & connections
Understand the moving parts.
STT
AssemblyAI · Deepgram · OpenAI · Google
LLM
OpenAI · Anthropic · Google Gemini
TTS
Cartesia · Deepgram · Fish Audio · Inworld · ElevenLabs · Rime · Speechify
TELEPHONY
LiveKit Phone Numbers · Third-party SIP
Integrations
Provider plugins · MCP (Python)
Evidence you can inspect.
A documented feature, a supplier claim and an unanswered question are different kinds of evidence.
Evidence coverage details
- Strong-source resolved
- 50
- Vendor-claim only
- 3
- Conditional
- 40
- Other sourced outcomes
- 0
- Unknown after research
- 21
- Conflicting evidence
- 0
- Not applicable
- 0
- Not researched
- 0
Inspect detailed evidence, sources and research gaps
September 2026 remains the historical baseline. A new research date does not mean a feature launched that day. Unknown after research and not researched are separate states.
Official documentation establishes what the supplier documents. No paid calls, independent product benchmarks or compliance audits were performed in this research pass. Coverage measures research work, never product quality.
The operating decision.
Price structure
Component stack
V3.3 evidence review · Cloud Ship starting subscription; session overage, inference, transport and artifacts extra
Plan plus session, telephony, inference, observability and other usage; example calculator minute total is not fixed price.
Human fallback
Unknown — no established fallback mode
No individually verified capability establishes this selection. Conditional support and vendor claims do not satisfy Strict Must-have.
Documented transfer is not a staffed human service. Verify who answers, when they answer and what happens if they do not.
Try it yourself
Unknown — no independently checked public demo entry is attached.
External signals
14,374 Repository stars
External signal — not a VoiceAgent.best score.
Repository interest at capture time, not installed users, revenue, reliability or a VoiceAgent.best score.
GitHubUnderlying model context: the documented stack above does not establish an exact benchmark checkpoint or its deployed settings. Public model results cannot become a LiveKit score.
Append-only price history
Every capture leaves a trace.
A new check creates a new snapshot. Deltas appear only for the same currency, unit, model and scope.
Tracked: $0.01 / minute · Agent-session overage only; excludes end-to-end call cost
Tracked: $50 / month · V3.3 evidence review · Cloud Ship starting subscription; session overage, inference, transport and artifacts extra
Tracked: $50 / month · V3.3 evidence review · Cloud Ship starting subscription; session overage, inference, transport and artifacts extra
New research observations preserve earlier captures. “Evidence improved” records what we learned; it does not claim the market changed on that date. September 2026 is the historical baseline.
Price the actual workflow.
Complete cost cannot be established from this reference. The calculator example $0.0479/min is assumption-specific; no maximum complete bill from minutes alone. · Inference, phone number/transport, recording and observability, outbound carrier, bandwidth and optional private links
Carry the current source into your estimate, then enter the model, voice, carrier and fixed fees your configuration needs.
Alternatives by category
Other developer frameworks to inspect.
Alphabetical alternatives within the same category. An editorial starting point, not a performance ranking.
Custom agents that need direct realtime model access and tool orchestration.
CompareTeams that want to own pipeline code and choose deployment infrastructure.
CompareEvidence archive
Inspect the record.
LiveKit Cloud pricingreviewed
Build is $0/month; Ship starts at $50/month; Scale starts at $500/month. Agent-session overage on paid plans is $0.01/minute. Models, telephony and observability have separate meters. Scale includes access to security reports and HIPAA arrangements.
Inspect official source ↗LiveKit Agents introductionreviewed
LiveKit documents an agents framework for realtime voice applications with model integrations and deployment options.
Inspect official source ↗LiveKit model integrationsreviewed
LiveKit Inference and provider plugins connect speech recognition, language models and speech synthesis; providers include OpenAI, Google, AssemblyAI, Deepgram, Cartesia and ElevenLabs.
Inspect official source ↗LiveKit agent-assisted transferreviewed
Warm transfer can hold the caller, dial a manager, provide context privately, connect both people and remove the agent. An unavailable manager can return control to the agent.
Inspect official source ↗LiveKit call forwardingreviewed
Cold transfer uses SIP REFER through a configured trunk and ends the LiveKit session when the caller leaves. Carrier support and trunk configuration are required.
Inspect official source ↗LiveKit MCP toolsreviewed
The Python Agents SDK supports MCP toolsets over HTTP and local stdio. The documentation labels this feature Python-only.
Inspect official source ↗Evidence describes documented scope, not a tested platform score. Unknown means a conclusion has not been established; API connectivity does not establish a native integration. Read our research methodology →
Public affiliate terms and an approved tracking link have not been verified.
Partner eligibility and commercial terms are unverified.