Documentation

Sign up for VoxaHub and verify your email. You get a free tier with limited usage to explore the platform.

Click "Create New Agent" from the dashboard. Choose a domain (Sales, Healthcare, Support, etc.), configure the personality, set up the voice, and add knowledge base content.

Use the built-in chat interface to test your agent with text or voice. Fine-tune the system prompt and personality traits until you are satisfied.

Deploy your agent and use the Integrations page to get a share link, embed widget code, or API endpoints to integrate into your website or application.

VoxaHub uses ElevenLabs for high-quality voice synthesis. Choose from a library of voices, select gender, and fine-tune speed and stability settings.

Voice input is automatically transcribed using ElevenLabs STT. The always-on microphone feature uses Voice Activity Detection (VAD) to auto-detect speech.

Enable the always-on mic mode for hands-free conversation. The agent listens continuously, detects when you speak, records your message, and responds automatically. Click the radio icon (📡) in the chat footer to toggle.

Upload PDF, DOCX, TXT, CSV, or JSON files. Add URLs for the agent to learn from. Create FAQ pairs for common questions. Add custom text instructions.

Knowledge base content is automatically extracted and injected into the agent's context when conversations happen. The agent uses this information to provide accurate, domain-specific responses.

Enable public access on your agent and share a direct link. Anyone with the link can chat with your agent without an account.

Copy the JavaScript snippet and add it to any website. A floating chat button appears in the bottom-right corner that opens a full chat interface in an iframe.

Use the API endpoints to build custom integrations. Send POST requests to the chat endpoint with a message and get responses with text and optional audio.

Configure webhook URLs to receive real-time notifications for events like conversation started, message received, message sent, and conversation ended.

Send a JSON body with { "message": "your text", "conversation_id": "optional" }. Returns { "text": "agent response", "audio_url": "optional TTS audio", "conversation_id": "id" }.

Send a multipart/form-data with an "audio" file field and optional "conversation_id". Returns the agent's text response, audio URL, and transcription of the user's speech.