Voxera SDK — Core

Core TypeScript SDK for Voxera Voice Platform. Platform-agnostic client library that powers all framework-specific SDKs (React, React Native, iOS, Android, Flutter).

Platform Endpoints

| Service | URL | |---------|-----| | Media Server (WebSocket) | wss://media.voxera-voice.com | | Demo | https://demo.voxera-voice.com |

Pass the Media Server URL as serverUrl when creating a client. Get your API key at app.voxera-voice.com.

Meeting Modes

AI Meeting (ai-meeting)

A real-time voice conversation with an AI assistant. The AI listens via speech-to-text (STT), processes through a configurable LLM (OpenAI, Anthropic, Ollama), and responds with natural speech via text-to-speech (TTS). Supports multiple participants talking to the same AI in a shared room.

Capabilities:

• Real-time voice AI conversation (bidirectional audio over WebRTC/mediasoup)
• Configurable AI persona via dashboard Agent settings (system prompt, model, temperature, max tokens)
• Multiple TTS providers (OpenAI, ElevenLabs, Azure) with voice selection
• Multiple STT providers (Google, OpenAI Whisper)
• Live transcription of all speakers
• Video and screen sharing with optional AI vision (enableVideoAI)
• Ask AI — send the current transcript to AI and get a spoken response
• Text message input alongside voice
• AI-generated meeting summaries with action items and key topics
• AI-generated meeting minutes (title, attendees, duration, sections)
• Bookmarks — mark key moments and flag as action items

Normal Meeting (normal-meeting)

A multi-participant audio/video meeting without AI voice. Participants talk to each other via WebRTC. AI features are available in text-only form.

Capabilities:

• Multi-participant real-time audio and video (WebRTC/mediasoup SFU)
• Host controls — mute/unmute participant, mute all, remove participant, transfer host, lock room, end meeting
• Waiting room — enable/disable, admit or deny individual participants, admit all
• Live transcription (toggle on/off by host)
• Ask AI (text) — send a prompt to AI, receive a streamed text response (no voice)
• AI-generated meeting summaries and minutes
• Bookmarks
• Screen sharing

Common Features (Both Modes)

• WebRTC via mediasoup — SFU-based low-latency audio/video transport with TURN fallback
• Room management — create rooms with codes, join by code, display names, participant list
• Real-time events — participant join/leave, mute status, host changes, room lock, transcription entries
• Webhook integration — receive server-side events on session start/end, messages, errors
• Usage tracking — minutes consumed, session history, breakdowns by period
• Configurable — override AI provider, voice, language, video resolution per session

Installation

npm install @voxera/sdk-core

Quick Start

import { MayaVoiceClient } from '@voxera/sdk-core';

const client = new MayaVoiceClient({
  appKey: 'your-app-key',                    // from https://app.voxera-voice.com
  serverUrl: 'wss://media.voxera-voice.com',
  agentId: 'your-agent-id',                  // created in the dashboard
  onConnectionStatusChange: (status) => console.log('Connection:', status),
  onMessage: (msg) => console.log('Message:', msg),
  onTranscript: (text, isFinal) => console.log('Transcript:', text),
});

// Connect and start AI conversation
await client.connect();
await client.startConversation();

// Later
await client.endConversation();
client.disconnect();

Multi-Participant Room

// Host creates a room
await client.connect();
const room = await client.createRoom('Alice', 'Team Standup', 'normal-meeting');
console.log('Room code:', room.roomCode); // share this code

// Another user joins
await otherClient.connect();
await otherClient.joinRoom(room.roomCode, 'Bob');

// Host controls
await client.muteParticipant(sessionId, targetClientId);
await client.lockRoom(sessionId, true);
await client.toggleTranscription(sessionId, true);

// AI features (text-only in normal-meeting mode)
await client.askAiText(sessionId, 'Summarize what was discussed');
await client.generateSummary(sessionId);
await client.generateMinutes(sessionId);
await client.addBookmark(sessionId, 'Key decision made', true);

Configuration

MayaVoiceConfig

| Field | Type | Required | Description | |-------|------|----------|-------------| | appKey | string | Yes | API key from the dashboard | | serverUrl | string | Yes | wss://media.voxera-voice.com | | agentId | string | No | Agent ID (created in the dashboard). Defines AI persona, model, voice, tools, and endpoints | | userId | string | No | Identify the end-user for analytics | | metadata | Record<string, string> | No | Custom key-value pairs passed to agent endpoints | | videoConfig | VideoConfig | No | Camera/video settings | | screenShareConfig | ScreenShareConfig | No | Screen share settings | | connectionOptions | ConnectionOptions | No | Reconnect/timeout settings |

VideoConfig

| Field | Type | Default | Description | |-------|------|---------|-------------| | enabled | boolean | false | Enable video on connect | | width | number | 640 | Video width | | height | number | 480 | Video height | | frameRate | number | 30 | Frames per second | | facingMode | string | user | user (front) · environment (back) | | enableVideoAI | boolean | false | Send video frames to AI pipeline |

ConnectionOptions

| Field | Type | Default | Description | |-------|------|---------|-------------| | autoReconnect | boolean | true | Auto-reconnect on disconnect | | reconnectAttempts | number | 5 | Max reconnect attempts | | reconnectDelay | number | 1000 | Delay between attempts (ms) | | timeout | number | 10000 | Connection timeout (ms) | | iceServers | RTCIceServer[] | — | Custom ICE/TURN servers |

Events

| Event | Payload | Description | |-------|---------|-------------| | connection:status | ConnectionStatus | idle → connecting → connected → disconnected | | conversation:status | ConversationStatus | idle → starting → active → ending → ended | | speaking:status | SpeakingStatus | none · user · ai | | message | ConversationMessage | New chat message | | transcript | (text, isFinal) | Real-time speech transcription | | participant:joined | RoomParticipant | Someone joined the room | | participant:left | RoomParticipant | Someone left the room | | participants:updated | RoomParticipant[] | Full participant list update | | you:muted | — | You were muted by the host | | you:removed | — | You were removed from the room | | host:changed | — | Host was transferred | | meeting:ended | — | Meeting ended by host | | room:locked | boolean | Room lock status changed | | transcription:toggled | boolean | Transcription toggled by host | | transcription:live | TranscriptionEntry | Live transcription entry | | ask-ai:started | — | AI voice response started | | ask-ai:processing | — | AI is processing | | ask-ai:cancelled | — | AI voice response cancelled | | ask-ai-text:started | — | AI text response started | | ask-ai-text:chunk | string | Streamed AI text chunk | | ask-ai-text:response | string | Complete AI text response | | summary:generated | MeetingSummary | AI summary ready | | minutes:generated | MeetingMinutes | AI minutes ready | | bookmark:added | MeetingBookmark | Bookmark created | | bookmark:removed | string | Bookmark removed (ID) | | waiting-room:updated | WaitingRoomEntry[] | Waiting room list changed | | waiting-room:toggled | boolean | Waiting room enabled/disabled |

Methods

Connection

| Method | Description | |--------|-------------| | connect() | Connect to the media server | | disconnect() | Disconnect and clean up | | connectSocketOnly() | Connect socket without WebRTC (for room setup) | | setupRoomWebRTC() | Initialize WebRTC after joining a room |

Conversation

| Method | Description | |--------|-------------| | startConversation() | Start voice conversation | | endConversation() | End voice conversation | | sendMessage(content) | Send a text message |

Media

| Method | Description | |--------|-------------| | setMuted(muted) | Mute/unmute microphone | | enableVideo() | Start camera | | disableVideo() | Stop camera | | toggleVideo() | Toggle camera on/off | | startScreenShare() | Start screen sharing | | stopScreenShare() | Stop screen sharing | | toggleScreenShare() | Toggle screen sharing | | getStats() | Get WebRTC connection stats |

Host Controls (requires host role)

| Method | Description | |--------|-------------| | muteParticipant(sessionId, targetId) | Mute a participant | | muteAll(sessionId) | Mute all participants | | unmuteAll(sessionId) | Unmute all participants | | removeParticipant(sessionId, targetId) | Remove a participant | | lockRoom(sessionId, locked) | Lock/unlock the room | | endMeeting(sessionId) | End the meeting for everyone | | transferHost(sessionId, targetId) | Transfer host role | | toggleTranscription(sessionId, enabled) | Enable/disable transcription | | enableWaitingRoom(sessionId, enabled) | Enable/disable waiting room | | admitParticipant(sessionId, targetId) | Admit from waiting room | | denyParticipant(sessionId, targetId) | Deny from waiting room | | admitAll(sessionId) | Admit all waiting participants |

AI Features

| Method | Description | |--------|-------------| | askAi(sessionId) | Ask AI to respond with voice (ai-meeting) | | cancelAskAi(sessionId) | Cancel ongoing AI voice response | | askAiText(sessionId, prompt?) | Ask AI for a streamed text response | | generateSummary(sessionId) | Generate AI meeting summary | | generateMinutes(sessionId) | Generate AI meeting minutes | | addBookmark(sessionId, label, isActionItem?) | Add a bookmark | | removeBookmark(sessionId, bookmarkId) | Remove a bookmark | | getBookmarks(sessionId) | Get all bookmarks | | getTranscript(sessionId) | Get full transcript | | getSummaries(sessionId) | Get all summaries | | getMinutes(sessionId) | Get all minutes |

License

MIT