What is a simulcast video API?

A simulcast video API is a programmable interface that enables publishers to send multiple quality layers of the same video stream simultaneously. The Selective Forwarding Unit (SFU) on the server selects which quality layer to forward to each subscriber based on their available bandwidth, viewport size, and whether the publisher is the active speaker. This approach avoids server-side transcoding while delivering optimal quality to every participant.

How much bandwidth does simulcast save?

Simulcast typically reduces server egress bandwidth by 60-80% compared to sending the highest quality layer to all subscribers. In a 20-person meeting, most participants appear as small thumbnails and receive the 180p layer instead of 720p. Only the active speaker and pinned participants receive the full 720p stream. This dramatically reduces CDN and bandwidth costs.

Does V100 simulcast require transcoding?

No. V100's simulcast implementation is zero-transcode. Publishers encode and send all three quality layers directly from their device. The SFU selects and forwards layers without re-encoding, which eliminates server-side CPU costs and reduces latency compared to transcoding-based approaches.

SIMULCAST

Simulcast Video API with
SFU Quality Selection

A simulcast video API is a programmable interface where publishers send three simultaneous quality layers (720p, 360p, 180p) and a Selective Forwarding Unit (SFU) chooses which layer to forward to each subscriber based on their available bandwidth, viewport size, and speaker status. V100's simulcast implementation reduces server egress by 60-80% with zero transcoding, automatic quality adaptation, and per-subscriber layer selection -- delivering optimal video quality to every participant without server-side re-encoding.

Get API Key View Integration Guide

3 quality layers per publisher

60-80% egress reduction

Zero server-side transcoding

Why Simulcast Matters for Video Conferencing

Without simulcast, every subscriber receives the same quality stream regardless of whether they are viewing a full-screen speaker or a 100-pixel thumbnail. This wastes bandwidth, increases costs, and degrades the experience for participants on constrained connections.

Bandwidth Inequality Is Real

In a typical 20-person meeting, participants range from fiber connections at 100+ Mbps to mobile hotspots at 1-2 Mbps. Without simulcast, the server either sends 720p to everyone (overwhelming mobile users) or downgrades everyone to the lowest common denominator. Simulcast sends each subscriber the best quality their connection can handle.

Egress Costs Dominate

Server egress bandwidth is the largest variable cost in video conferencing infrastructure. In a 20-person meeting without simulcast, sending 720p to all subscribers costs 20x the egress of a single stream. With simulcast, 15 thumbnail viewers get 180p (16x smaller) while only the 5 visible speakers get 720p -- reducing total egress by 60-80%.

Transcoding Is Expensive

The alternative to simulcast is server-side transcoding, which decodes and re-encodes video at multiple quality levels on the SFU. This requires expensive GPU infrastructure, adds 100-500ms latency per transcode, and is the primary scaling bottleneck for video platforms. V100's simulcast eliminates transcoding entirely.

How V100 Simulcast Works

V100's simulcast pipeline operates in three stages: publish, select, and forward. Publishers encode three layers locally. The SFU selects the optimal layer for each subscriber. No video is decoded or re-encoded on the server.

Publish 3 Layers

Each publisher's browser encodes and sends three simultaneous quality layers to the SFU:

HIGH 720p at 1.5 Mbps -- full quality for active speakers and pinned views

MID 360p at 500 Kbps -- sidebar and gallery views on desktop

LOW 180p at 150 Kbps -- thumbnails, mobile, and constrained connections

SFU Selects Per Subscriber

The SFU evaluates three signals for each subscriber to determine which layer to forward:

Bandwidth: Available downstream capacity

Viewport: Tile size on the subscriber's screen

Speaker: Active speaker gets priority for high layer

Layer switching happens seamlessly. Subscribers never see a quality change notification -- the SFU switches on keyframe boundaries for smooth transitions.

Forward Without Transcoding

The SFU forwards the selected layer directly to each subscriber without decoding or re-encoding. This zero-transcode approach eliminates:

Server-side GPU/CPU encoding costs

Transcoding latency (100-500ms saved)

Quality degradation from re-encoding

Scaling bottleneck at 50+ participants

Egress Savings: 20-Person Meeting

Scenario	Per-Subscriber	Total Egress	Savings
No simulcast (720p to all)	19 x 1.5 Mbps	28.5 Mbps	--
V100 simulcast	4x720p + 15x180p	8.25 Mbps	71% reduction

Integration Guide

Simulcast is enabled by default on all V100 meetings. Publishers automatically send 3 quality layers and the SFU handles selection. You can also control layer selection programmatically.

simulcast-config.js

// npm install v100-sdk
import { V100 } from 'v100-sdk';

const v100 = new V100('YOUR_API_KEY');

// Create a meeting with simulcast enabled (default)
const session = await v100.meetings.create({
  name: 'Team Standup',
  simulcast: {
    enabled: true,              // Default: true
    layers: [
      { rid: 'high',  width: 1280, height: 720,  maxBitrate: 1_500_000 },
      { rid: 'mid',   width: 640,  height: 360,  maxBitrate: 500_000 },
      { rid: 'low',   width: 320,  height: 180,  maxBitrate: 150_000 }
    ]
  }
});

// Programmatically control which layer a subscriber receives
const meeting = await v100.meetings.join('mtg_abc123');

// Pin a participant to high quality (e.g., they're presenting)
meeting.setSubscriberLayer('participant_xyz', 'high');

// Get real-time quality stats per subscriber
meeting.on('qualityChange', (event) => {
  console.log(`${event.participantId}: ${event.layer}`);
  // → "participant_xyz: high"
  // → "participant_abc: low"
});

// Monitor bandwidth and layer statistics
const stats = await meeting.getSimulcastStats();
// stats.totalEgress      → 8250000 (bits/sec)
// stats.savedVsNoSimulcast → "71%"
// stats.layerDistribution → { high: 4, mid: 1, low: 15 }

Or configure simulcast via the REST API:

terminal

curl -X POST https://api.v100.ai/v1/meetings \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Team Standup",
    "simulcast": {
      "enabled": true,
      "layers": [
        { "rid": "high", "width": 1280, "height": 720, "maxBitrate": 1500000 },
        { "rid": "mid",  "width": 640,  "height": 360, "maxBitrate": 500000 },
        { "rid": "low",  "width": 320,  "height": 180, "maxBitrate": 150000 }
      ]
    }
  }'

Use Cases

Simulcast benefits every video conferencing scenario, from small team calls to large-scale events with hundreds of participants on diverse network conditions.

Large Meetings (20-200)

In gallery views with 20+ participants, most tiles are small thumbnails. Simulcast sends 180p to thumbnails and 720p only to the active speaker, reducing total egress by 70%+ while maintaining visual clarity where it matters.

Mobile Participants

Mobile users on cellular connections receive the low or mid layer automatically, preventing buffering and dropped frames. When a mobile user goes full-screen on the active speaker, the SFU upgrades them to the high layer if bandwidth permits.

Global Teams

Teams with members across regions face varying network quality. Simulcast ensures participants in bandwidth-constrained regions receive a smooth low-quality stream rather than a stuttering high-quality one, while high-bandwidth participants get full resolution.

Pricing

Simulcast is included on all V100 plans at no additional cost. You are billed for video minutes, not for the number of quality layers. Simulcast actually reduces your costs by lowering total bandwidth consumption.

Free Tier

$0/mo

Simulcast included.

3 quality layers

Automatic layer selection

60 minutes/month

Quality stats API

RECOMMENDED

Pro

Usage-based

Full simulcast control.

$0.002/min video

Custom layer configuration

Programmatic layer pinning

Real-time analytics

Enterprise

Custom

Dedicated SFU infrastructure.

Volume discounts

Dedicated media servers

Custom SFU policies

Priority support

Reduce Video Egress 60-80% with Simulcast

Simulcast is included on all plans. Zero transcoding, automatic quality adaptation, per-subscriber layer selection. Free tier available.

Get API Key — Free Tier Contact Sales

Related Resources

Feature

Simulcast Video API with
SFU Quality Selection

Why Simulcast Matters for Video Conferencing

Bandwidth Inequality Is Real

Egress Costs Dominate

Transcoding Is Expensive

How V100 Simulcast Works

Publish 3 Layers

SFU Selects Per Subscriber

Forward Without Transcoding

Egress Savings: 20-Person Meeting

Integration Guide

Use Cases

Large Meetings (20-200)

Mobile Participants

Global Teams

Pricing

Reduce Video Egress 60-80% with Simulcast

Related Resources

Adaptive Bitrate API

Active Speaker Detection

Noise Suppression API

Simulcast Video API withSFU Quality Selection

Why Simulcast Matters for Video Conferencing

Bandwidth Inequality Is Real

Egress Costs Dominate

Transcoding Is Expensive

How V100 Simulcast Works

Publish 3 Layers

SFU Selects Per Subscriber

Forward Without Transcoding

Egress Savings: 20-Person Meeting

Integration Guide

Use Cases

Large Meetings (20-200)

Mobile Participants

Global Teams

Pricing

Reduce Video Egress 60-80% with Simulcast

Related Resources

Adaptive Bitrate API

Active Speaker Detection

Noise Suppression API

Simulcast Video API with
SFU Quality Selection