Skip to content
AI Voice Agents
AI Voice Agents10 min0 views

WebRTC + AI for Smart Home Voice Control in 2026: Matter 1.5.1 Cameras and Two-Way Audio

Matter 1.5.1 made WebRTC the official spec for smart-home cameras and two-way audio. Aqara's first Matter cameras shipped H1 2026. Here is the AI voice-control build pattern.

Matter 1.5.1 made WebRTC the official transport for smart-home cameras and two-way audio in 2026. Home Assistant 2026.1 added native WebRTC camera streaming. Aqara shipped the first Matter-certified camera in H1 2026. Voice control is the consumer-facing surface; WebRTC is the plumbing.

Why this matters

Smart-home voice through 2024 was Alexa, Google Home, and Siri — three closed ecosystems. Matter 1.5.1 (December 2025) standardized the interop, including cameras with two-way audio and WebRTC transport. Now any Matter-compliant voice agent can talk to any Matter-compliant camera, lock, doorbell, or thermostat.

For a CallSphere-style architecture, this matters because the same Pion Go gateway 1.23 that runs voice-AI for real-estate or healthcare can be a Matter Controller — exposing a single voice agent that controls every device on a home network. Josh.ai already does this; the open-source path is now realistic.

Architecture

```mermaid flowchart LR User[User Mic] -- WebRTC voice --> Gateway[Pion Go gateway 1.23] Gateway --> AI[Voice Agent + Tool Calls] AI --> Matter[Matter Controller] Matter -- IPv6 --> Camera[Matter Camera] Matter -- IPv6 --> Lock[Matter Lock] Camera -- WebRTC video+audio --> Gateway Gateway --> User ```

Hear it before you finish reading

Talk to a live CallSphere AI voice agent in your browser — 60 seconds, no signup.

Try Live Demo →

CallSphere implementation

CallSphere does not ship a smart-home product, but the pattern reuses three core components from the platform:

  • Real Estate (OneRoof) listing tours — Listing agents demo Matter-compatible properties through the same WebRTC pipeline; buyers see live camera feeds and the agent can control lights and locks for showings. Pion Go gateway 1.23 + NATS + 6-container pod (CRM, MLS, calendar, SMS, audit, transcript). See /industries/real-estate.
  • /demo — The marketing demo includes a Matter-style "control my smart home" toggle that demonstrates a single voice agent controlling a virtual home. Try it at /demo.
  • Healthcare — HIPAA-respecting integration with smart-home aging-in-place devices (alerts, medication reminders) over the same gateway.

37 agents, 90+ tools, 115+ tables, 6 verticals, HIPAA + SOC 2. $149/$499/$1499 pricing; 14-day /trial; 22% /affiliate.

Build steps with code

```typescript // 1. WebRTC voice agent that maps tool calls to Matter clusters import { OpenAIRealtime } from "@callsphere/realtime"; import { MatterController } from "@callsphere/matter"; const matter = new MatterController({ fabric: "home-1234" }); const agent = new OpenAIRealtime({ tools: [ { name: "set_light", description: "Turn a light on/off or set brightness", parameters: { roomId: "string", on: "boolean", level: "number" }, handler: async ({ roomId, on, level }) => { const node = await matter.findByRoom(roomId, "OnOffLight"); await node.cluster("OnOff").write({ onOff: on }); if (level) await node.cluster("LevelControl").write({ currentLevel: level }); }, }, { name: "stream_camera", description: "Stream a Matter camera to the user's browser via WebRTC", parameters: { cameraId: "string" }, handler: async ({ cameraId }) => { const cam = await matter.find(cameraId); const offer = await cam.cluster("WebRTCTransport").requestOffer(); return await rtc.bridgeToUser(offer); }, }, ], });

// 2. Matter 1.5.1 camera WebRTC bridge import { Camera } from "@matter/camera"; const cam = new Camera({ commissioner: matter }); cam.on("offer", async (offer) => { const answer = await pcUserSide.setRemoteAndAnswer(offer); cam.sendAnswer(answer); }); ```

Pitfalls

  • Trying to go peer-to-peer browser-to-camera — most home cameras live behind CGNAT; force TURN.
  • Voice agent that controls devices without confirmation — "lock the door" tool calls need a confirm step.
  • Local-network-only assumptions — Matter supports remote control via Hub+Cloud; design for both.
  • Over-permissive tool scopes — bound the agent to a room or zone; never give whole-home control to a voice prompt without auth.
  • Latency on commands — Matter UDP commands round-trip in 50-200 ms locally; voice agent should TTS confirmation only after the device acks.

FAQ

Does Matter replace Alexa/Google? No — they support Matter as one of many ecosystems; Matter is interop, not UI.

Still reading? Stop comparing — try CallSphere live.

CallSphere ships complete AI voice agents per industry — 14 tools for healthcare, 10 agents for real estate, 4 specialists for salons. See how it actually handles a call before you book a demo.

Is WebRTC mandatory for Matter cameras? Per spec 1.5.1, WebRTC is the standardized transport for two-way audio + video.

Can I deploy this without a Matter controller? No — you need a Matter-certified controller (or a self-hosted one like Home Assistant + python-matter-server).

How do I handle multi-user homes? Per-user voice biometrics + per-user auth scope on the agent's tool calls.

Privacy concerns? Camera feeds stay on-network unless explicitly remoted; voice agent transcripts respect HIPAA-like local retention.

Sources

Try a Matter-style demo at /demo, see /pricing, or /trial.

Share

Try CallSphere AI Voice Agents

See how AI voice agents work for your industry. Live demo available -- no signup required.

Related Articles You May Like