--- title: Build Zoom, Twitch, or Voice AI Using Just 3 Lego Blocks description: LiveKit's architecture is built on three universal primitives — Room, Participant, and Track — that can be reconfigured to build any real-time application,... canonical: https://daily.dev/posts/build-zoom-twitch-or-voice-ai-using-just-3-lego-blocks-liii1i5xk twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Build Zoom, Twitch, or Voice AI Using Just 3 Lego Blocks | daily.dev og:description: LiveKit's architecture is built on three universal primitives — Room, Participant, and Track — that can be reconfigured to build any real-time application,... og:url: https://daily.dev/posts/build-zoom-twitch-or-voice-ai-using-just-3-lego-blocks-liii1i5xk og:image: https://api.daily.dev/og/posts/liIi1I5XK.png og:image:alt: Build Zoom, Twitch, or Voice AI Using Just 3 Lego Blocks og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Build Zoom, Twitch, or Voice AI Using Just 3 Lego Blocks **[Medium](https://daily.dev/sources/medium_js)** · 8 min read · 0 upvotes · 0 comments ## Summary LiveKit's architecture is built on three universal primitives — Room, Participant, and Track — that can be reconfigured to build any real-time application, from Zoom-like video calls to Twitch-style live streams to Voice AI agents. Rooms act as containers with shared metadata, Participants can be humans, AI agents, IoT devices, or phone callers, and Tracks are individual audio or video streams. The platform uses a Pub/Sub model via an SFU (Selective Forwarding Unit) to route media efficiently, with Simulcast for adaptive video quality and Dynacast to dynamically pause unused video layers and save bandwidth. Audio quality is handled automatically via Opus codec and RED (Redundant Audio Data) for packet loss recovery. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What are the core building blocks in LiveKit for building real-time apps like Zoom or Twitch clones? LiveKit is built around three primitives: Room, Participant, and Track. A Room is the universal container holding shared mutable metadata; a Participant is any entity connected to the room (human, AI agent, IoT camera, or drone) with an identity, name, and attributes; a Track is a single audio or video stream. Combining these differently produces video calls, live streams, or voice AI apps without changing the underlying stack. _daily.dev surfaces deep dives like this for developers architecting real-time voice or video features._ ### What is the difference between LiveKit Room Metadata and Participant Attributes? Room Metadata is a shared key-value store attached to the room itself and persists as long as the room stays open, even if individual users disconnect. Participant Attributes are key-value pairs attached to a specific person and are instantly synchronized with everyone in the room, but they vanish immediately when that participant disconnects or loses connection. _track these kinds of state-management distinctions on daily.dev when designing multi-agent or live-room apps._ ### How does Dynacast improve on Simulcast for bandwidth optimization in WebRTC video calls? Dynacast makes the SFU tell a publishing device to stop sending video quality layers nobody is subscribed to, instead of Simulcast's default behavior of always uploading all three layers (high, medium, low) regardless of demand. When a subscriber requests a higher layer again, the SFU signals the device to resume sending it, saving battery and upload bandwidth on the publishing side. _developers tuning video bandwidth costs can follow SFU optimization techniques like this on daily.dev._ --- Tags: [#webrtc](https://daily.dev/tags/webrtc), [#voice-ai](https://daily.dev/tags/voice-ai) [View this post on daily.dev](https://daily.dev/posts/build-zoom-twitch-or-voice-ai-using-just-3-lego-blocks-liii1i5xk) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Build Zoom, Twitch, or Voice AI Using Just 3 Lego Blocks","url":"https://daily.dev/posts/build-zoom-twitch-or-voice-ai-using-just-3-lego-blocks-liii1i5xk","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/build-zoom-twitch-or-voice-ai-using-just-3-lego-blocks-liii1i5xk"},"datePublished":"2026-07-26T05:47:25.303Z","dateModified":"2026-09-14T08:03:11.840Z","description":"LiveKit's architecture is built on three universal primitives — Room, Participant, and Track — that can be reconfigured to build any real-time application,...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/101d46fef88a11cc59ebdea15b842a44?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/101d46fef88a11cc59ebdea15b842a44?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Medium","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Medium","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/medium","url":"https://daily.dev/sources/medium_js"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/build-zoom-twitch-or-voice-ai-using-just-3-lego-blocks-liii1i5xk","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"webrtc,voice-ai","timeRequired":"PT8M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Medium","item":"https://daily.dev/sources/medium_js"},{"@type":"ListItem","position":3,"name":"Build Zoom, Twitch, or Voice AI Using Just 3 Lego Blocks"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/build-zoom-twitch-or-voice-ai-using-just-3-lego-blocks-liii1i5xk#faq","mainEntity":[{"@type":"Question","name":"What are the core building blocks in LiveKit for building real-time apps like Zoom or Twitch clones?","acceptedAnswer":{"@type":"Answer","text":"LiveKit is built around three primitives: Room, Participant, and Track. A Room is the universal container holding shared mutable metadata; a Participant is any entity connected to the room (human, AI agent, IoT camera, or drone) with an identity, name, and attributes; a Track is a single audio or video stream. Combining these differently produces video calls, live streams, or voice AI apps without changing the underlying stack. daily.dev surfaces deep dives like this for developers architecting real-time voice or video features."}},{"@type":"Question","name":"What is the difference between LiveKit Room Metadata and Participant Attributes?","acceptedAnswer":{"@type":"Answer","text":"Room Metadata is a shared key-value store attached to the room itself and persists as long as the room stays open, even if individual users disconnect. Participant Attributes are key-value pairs attached to a specific person and are instantly synchronized with everyone in the room, but they vanish immediately when that participant disconnects or loses connection. track these kinds of state-management distinctions on daily.dev when designing multi-agent or live-room apps."}},{"@type":"Question","name":"How does Dynacast improve on Simulcast for bandwidth optimization in WebRTC video calls?","acceptedAnswer":{"@type":"Answer","text":"Dynacast makes the SFU tell a publishing device to stop sending video quality layers nobody is subscribed to, instead of Simulcast's default behavior of always uploading all three layers (high, medium, low) regardless of demand. When a subscriber requests a higher layer again, the SFU signals the device to resume sending it, saving battery and upload bandwidth on the publishing side. developers tuning video bandwidth costs can follow SFU optimization techniques like this on daily.dev."}}]} ```