Skip to main content

Node Overview

Every element on the Voiceblox canvas is a node. Nodes are grouped into four categories.

Framework

The Framework node is the central orchestrator. There is exactly one per flow.
NodeDescription
LiveKitRuns the agent on LiveKit’s real-time infrastructure

Component

Component nodes define the AI capabilities of your agent.
NodeHandleDescription
LLMllm_inLanguage model for generating responses
TTStts_inText-to-speech for voice output
STTstt_inSpeech-to-text for voice input
Avataravatar_inVisual avatar rendering (optional)
Personapersona_inAgent name, system prompt, and language

Conversation

Conversation nodes form the step chain — the sequence of actions during a call.
NodeDescription
StartEntry point. Agent speaks its opening line.
BurstA fixed number of back-and-forth exchanges.
Open TalkUnlimited free-form conversation on a topic.
TimerConversation capped at a time duration.
If/ElseBinary branch based on an LLM-evaluated condition.
CategorizeMulti-way branch by semantic classification.
TransferHand off to another agent.
SummarizeSummarize the conversation so far.
EndTerminates the conversation.

Integration

NodeHandleDescription
MCPtools_inExposes tools from an MCP server to the LLM
Exatools_inGives the LLM real-time web search
WebhookFires an HTTP POST at a conversation step
Vector StoreRetrieval-augmented generation (coming soon)

Connection diagram

[Persona] ──persona──▶ [Framework] ◀──llm── [LLM]
                           │         ◀──tts── [TTS]
                           │         ◀──stt── [STT]
                           │         ◀──tools── [MCP]

                       [Start] ──▶ [Burst] ──▶ [End]