arrows-left-rightAPI Terminology Mapping

Map between conceptual documentation names and API endpoint paths, SDK resources, and streaming event types.

The Amigo platform documentation uses conceptual names that are easier to understand, while the API uses internal system names. This page maps between them so you can connect what you read in the docs to what you see in the API.

Classic API Mappings

These mappings apply to the Classic API (REST endpoints and SDKs).

Conceptual Name
API Name
Description

Context Graph

service_hierarchical_state_machine

The structured map of states and transitions that guides agent behavior through a workflow.

Action

tool

A program or integration the agent can execute during a conversation (scheduling, data lookup, notifications).

Dynamic Behavior

dynamic_behavior_set

Runtime behaviors that activate based on conversational context or events.

Agent

agent

The agent definition including persona, background, and directives.

Metric

metric

A configured evaluation criterion for measuring agent performance.

Simulation

simulation

A test case combining a persona and scenario for automated conversation testing.

Persona

persona

A synthetic user profile used in simulation testing.

Scenario

scenario

A defined situation and context used in simulation testing.

Unit Test Set

unit_test_set

A group of simulation-based tests with success criteria.

Side-Effect State

ToolCallState

A context graph state that triggers external actions (data retrieval, API calls, alerts).

Decision State

DecisionState

A context graph state for routing and escalation decisions.

Action State

ActionState

A context graph state for user-facing responses.

Reflection State

ReflectionState

A context graph state for strategic reasoning.

Recall State

RecallState

A context graph state for memory integration.

Platform API Mappings

These mappings apply to the Platform API, which supports voice and workspace-based deployments.

Conceptual Name
API Name
Description

Workspace

workspace

A container that groups skills, agents, and configurations for a specific deployment.

Skill

skill

A discrete capability exposed by a workspace, such as scheduling or symptom assessment.

Monitor

monitor

A tracked signal or condition that can trigger alerts or behavior changes.

Operator

operator

A human staff member who receives escalated conversations with full handoff context.

World Model

world_model

The unified data layer that assembles information from EHR, conversations, and other sources.

Connector Runner

connector_runner

The integration layer that pulls external data into the platform.

Regulation Template

regulation_template

A compliance framework encoding regulatory requirements as agent constraints.

Confidence Score

confidence_score

A numeric measure of agent certainty used for escalation decisions.

Entity Resolution

entity_resolution

The process of matching records across data sources to a single identity.

Streaming Event Types

chevron-rightNDJSON streaming event typeshashtag

These event types appear in NDJSON streams from conversation endpoints.

Event
NDJSON Type
Description

Conversation created

conversation-created

Emitted when a new conversation starts.

Agent message chunk

new-message

A chunk of the agent's response, streamed as it is generated.

Interaction complete

interaction-complete

Emitted when the agent finishes processing a user message.

Dynamic behavior triggered

current-agent-action (subtype: dynamic-behavior-triggered)

A dynamic behavior activated during the conversation.

Tool execution started

current-agent-action (subtype: tool-call-started)

An action (tool) began executing.

Error

error

An error occurred during processing.

Version Set Presets

chevron-rightChannel-specific version set presetshashtag

Version set presets configure the agent's behavior profile for different channels.

Channel
Preset Name
Notes

Voice (standard)

voice

Max 5 states, action/annotation/tool-call types only

Voice (premium)

voice_premium

Same constraints, premium model

Voice (ultra-low latency)

voice_ultra_low

Single state only, no memory retrieval

Text

text

Unlimited states, all types

Text (premium)

text_premium

Same, premium model

Async

async

Unlimited states, dynamic behaviors

Async (premium)

async_premium

Same, premium model

circle-info

The OpenAPI spec at https://api.amigo.ai/v1/openapi.json is the definitive source for all endpoint paths, parameter names, and response schemas.

Last updated

Was this helpful?