AI text-to-speech
Natural-sounding audio prompts from text.
Voice agents that hold real conversations, live transcription that turns calls into usable data, and more - all running inside PBXware, managed from one hub, deployable in days, and running on infrastructure you already have.
AI Hub is the control center for AI across the platform. Connect your preferred providers, assign each to a specific job - text-to-speech, speech-to-text, or real-time speech - and manage it all per tenant. Each tenant uses its own key, with usage and spend limits set at your provider.
Connect and configure AI providers independently for each customer.
OpenAITTS · STT · STS
DeepgramSTT
ElevenLabsTTS · STS
GoogleSTT · STS
xAISTS
AWSSTT
Assign the right provider to text-to-speech, speech-to-text, or real-time speech.
Providers, agents, MCP servers and transcription, all in one menu.
Providers
Voice agents
MCP servers
Live transcription
Call transcription
Text to speech
Each tenant has its own key. Usage and spend limits live at your provider.
AI ships inside the platform rather than as a product you buy, license and integrate separately. That changes what it takes to start selling it, and what it costs you to keep running.
You bring your own AI provider account - we don't supply the LLM or tokens.
AI connects directly to live calls with a real-time speech-to-speech experience, natural, low-latency conversations, no extra servers or middleware. Just add your provider key and configure the agent.
When it can’t fully resolve a request, it transfers to a live agent, queue, department, extension, or a more specialized AI agent, forwarding the conversation context so nothing is repeated.
Connect an agent to CRMs, ticketing, calendars and knowledge bases straight from the prompt, so it can verify a caller, book an appointment or update a ticket on its own.
AI-handled calls are marked in CDR, with the transcript, event log and call details one click away. If a call moves between AI agents, you can review each agent's part separately.

"AI agents are here to make things easier for human agents by taking care of the routine jobs
nobody actually wants to do."
When a recorded call starts, Live Transcription streams the audio to your chosen speech-to-text provider and sends the text in real time to a WebSocket endpoint you define, ready for your CRM, helpdesk, analytics, a dashboard, or a custom app.
AgentThanks for calling, how can I help?
CallerI need to move my appointment.
AgentSure - which day works best?
Text arrives during the call, not after it ends.
Switched on per extension, queue or Enhanced Ring Group, rather than across the whole system.
Server-side on the platform, publishing wherever you point it.
The value isn't the transcript itself, it's what you connect it to. That layer is yours to build and sell.
A walkthrough of the two v8 capabilities: Voice Agents, which answer and handle calls, and Live Transcription, which streams the conversation as text to your systems.
Natural-sounding audio prompts from text.
Every voicemail, as readable text.
00:04AgentThanks for calling, how can I help?
00:09CallerI'd like to check my last invoice.
00:15AgentOf course, one moment please.
Recorded calls as text, with optional AI summaries.
Stereo recordings, ready for AI.
All included across Core, CX Lite & CX Pro.
Sell AI without buying a second platform.