mediasfu.com
├── Apps ──────── Web · Windows · macOS · Linux · iOS · Android
├── Widgets ───── 6 embeddable web components (one CDN script)
├── SDKs ──────── React · React Native · Flutter · Angular · Vue · Expo · Kotlin
├── Android ───── Native Kotlin SDK + mediasoup-client-android (WebRTC transport)
├── MediaSFUOpen ─ Self-host the full backend on your own infrastructure
├── AI Agents ─── Voice & vision pipelines (BYOK or managed)
├── Telephony ─── SIP trunking · PSTN · Cloud phone · IVR
└── Translation ─ 20+ languages · real-time · in-call

MediaSFU Cloud and MediaSFUOpen share the exact same SDKs. The live meeting experience at mediasfu.com/meeting is the same prebuilt UI you get out of the box with every SDK — fully customizable, unbranded, and ready to embed. MediaSFUOpen lets you run the entire backend on your own servers, with zero changes to your frontend code.

Add real-time communication to any website with a single script tag — no framework required.

<script src="https://cdn.mediasfu.com/widget.js"></script>

Example — add a call button in 3 lines:

<script src="https://cdn.mediasfu.com/widget.js"></script>
<mediasfu-call-button
  widget-key="wk_your_key"
  destination="+1-800-555-1234"
  theme="dark"
  button-text="Call Sales"
/>

Explore all widgets: mediasfu.com/widgets · Developer guide: mediasfu.com/developers

Build and deploy real-time AI agents powered by data buffers — no WebSocket plumbing required.

Audio In ──→ STT ──→ LLM ──→ TTS ──→ Audio Out
Video In ──→ Frame Analysis ──→ LLM ──→ Overlay / Response

Supported LLMs: ChatGPT · Claude · Gemini · DeepSeek · Llama · and more

Try now: agents.mediasfu.com · SDKs: github.com/MediaSFU/Agents

Production-ready, full-featured SDKs — each ships as a complete app you can customize or run headless.

MediaSFU provides a complete native Android solution built on Kotlin, with feature parity against the React and Flutter SDKs.

The mediasoup-client-android library handles the WebRTC signaling and media transport layer, while the MediaSFU Android SDK builds the full room experience on top of it — including participant grids, audio/video controls, polls, breakout rooms, and real-time translation.

Quick start monorepo with preconfigured apps for every SDK: MediaSFU-QuickStart-Apps

• HD Video Meetings — up to 3,000 participants
• Cloud Phone — SIP trunking, PSTN, IVR, call forwarding
• Screen Sharing with real-time annotation (web)
• Collaborative Whiteboards (web)
• Breakout Rooms — sub-sessions within a meeting
• Panellists & Audience Roles — separate presenter and viewer tiers with controlled access
• Polls — launch, respond to, and tally polls mid-session
• Virtual Backgrounds — blur, replace, or customize your video background
• Chat — direct & group messaging

• Cloud Recording — track-based, with watermarks, name tags, custom backgrounds
• Pause & Resume — unlimited, mid-session control
• Dual-View Recomposition — landscape and portrait from a single recording
• WebRTC Recorder — send media from anywhere; get real-time image/audio buffers

• Voice Pipelines — STT → LLM → TTS, fully customizable
• Vision Pipelines — real-time frame analysis with AI overlays
• Live Translation — 20+ languages, in-call, real-time; switch your speaking or listening language independently, like tuning to a channel
• Bring Your Own Keys — or use MediaSFU-managed keys

• 7 SDKs — React, Angular, Vue, Flutter, React Native, Expo, and native Kotlin for Android
• 6 Embeddable Widgets — one CDN script, works on any site
• Headless Mode — SDKs work without UI for custom integrations
• Prebuilt UI — unbranded, full-featured, ready to embed
• Managed Events — auto-handle idle participants, capacity limits, timeouts

Build a complete app with MediaSFU from scratch with our step-by-step tutorial repos:

No per-seat fees. No feature gates. Everything included at one rate.

• Community Forum: mediasfu.com/forums
• Twitter / X: @media_sfu
• YouTube: MediaSFU Channel
• GitHub Issues: Report a bug
• Website: mediasfu.com