Browser support
Which browsers can run a Synento session, how the SDK reports it, and what has actually been verified.
Browser support
Synento encodes in the browser with WebCodecs and ships frames over a WebSocket. There is no WebRTC fallback; that is a deliberate choice. Every frame is written to a durable stream, so each session is already its own recording. A browser either has the APIs or it cannot publish. The SDK checks before connecting and tells you which piece is missing.
Requirements
A browser needs the first four; the fifth is optional:
| Capability | Used for |
|---|---|
WebSocket | Transport for media, chat and control |
VideoEncoder / VideoDecoder / VideoFrame (WebCodecs) | Encoding your camera, decoding everyone else's |
navigator.mediaDevices | Camera and microphone capture (requires a secure context) |
MediaStreamTrackProcessor | Microphone capture, read straight from the track (Chrome, Edge). Where it is missing (Safari, Firefox), capture runs through AudioWorkletNode instead |
AudioWorkletNode | Playing the call's audio, and the microphone capture fallback |
AudioEncoder / AudioDecoder (Opus) | Optional. Compresses audio ~8x. Without it the call falls back to uncompressed PCM and works exactly the same, so this does not affect whether a browser is supported. checkSupport() reports it as webCodecsAudio for visibility only. |
navigator.mediaDevices only exists on HTTPS or localhost. On a plain
HTTP origin, including a LAN IP such as http://192.168.1.10:3000, capture is
unavailable no matter which browser you use.
Matrix
| Browser | Publish + subscribe | Notes |
|---|---|---|
| Chrome / Edge 94+ | Yes. Chromium is tested on every CI run | The reference target. The SDK prefers H.264, then VP9, then VP8. |
| Safari 16.4+ (macOS) | Expected; not yet verified on real Safari | WebCodecs video shipped in 16.4. The WebKit engine passes the full automated call (see below), but the shipping Safari has not been run yet. |
| Safari 16.4+ (iOS / iPadOS) | Expected, with caveats; not yet verified | No automated or manual iPhone/iPad run yet. See iOS below. |
| Firefox 133+ | Expected; not yet verified | 133 is the minimum Synento states; the SDK checks for the APIs, not a version number. No automated or manual Firefox run has succeeded yet. See below for why. |
| Older browsers without WebCodecs video | No | VideoEncoder is absent; the SDK reports it and refuses to connect. |
| Any browser over plain HTTP | No | No mediaDevices outside a secure context. |
"Expected" means the browser has every API the SDK checks for, not that a call has been seen to work in it. "Engine check" means an automated run on the browser's engine as Playwright builds it. Useful, but not the browser people actually install.
What has actually been verified
Being precise about this matters more than the table looking green.
-
Chromium: automated, every CI run. A two-party call with synthetic camera and microphone hardware: both sides decode the other's video, chat is delivered, and both participants' camera and microphone tracks are checked for recorded data in S2. See
e2e/tests/live-call.spec.ts. -
Chromium: A/V quality, run locally before media changes. Two generated talking heads hold a call, and what each hears and sees is measured against the original clips: lip-sync, dropouts, level, noise, frequency response, PESQ and STOI, live, in both replays and in the MP4 export. Not on CI, which hasn't the CPU for it. See
e2e/tests/av-quality.spec.tsanddocs/decisions.md("A/V quality is measured, not assumed"). Chromium only: no other engine here can use a file as its camera and microphone, so WebKit's capture and playback quality is unmeasured. -
Playwright WebKit: automated engine check, media included. WebKit in Playwright has mock camera and microphone devices, and on it:
- the recording spec publishes camera and mic, then replays and exports the session to an MP4 whose audio is audible, every CI run (Linux);
- the full two-party call (decoded video both ways, chat, camera and mic
recorded for both people) passed on macOS on 9 October 2026, 5 runs out of
5. It is opt-in (
E2E_WEBKIT_MEDIA=1 npm run test:e2e) until it has been tried on CI's Linux WebKit.
The engine reports H.264 encode and decode, so it sends H.264, and it has Opus for audio. (The MP4 export reads H.264, VP9 and VP8 cameras alike.) A WebKit pass is not a Safari pass. Playwright's WebKit is a build of the engine, not the shipping browser: its media stack, permissions and capture devices differ from Safari's. It narrows the risk; it does not remove it.
-
One camera per browser in WebKit. When two participants shared one WebKit browser, the second to start capturing silenced the first's camera: its preview went black and the far side never saw it. In separate browsers the call works. Safari behaves the same way across tabs, so do not test two participants in two tabs of one Safari. Use two devices, or Safari plus another browser.
-
Safari on macOS and iOS: manual, not yet run. Safari can be automated on a Mac (
safaridriver), but only after someone turns on Safari ▸ Settings ▸ Developer ▸ "Allow remote automation" (and runssafaridriver --enable, which asks for the Mac's password). It was off on the machine this was checked on, and even with it on, automation cannot grant camera access by itself. The checklist below is the real verdict; until it has been run, the Safari rows above stay "not yet verified". -
Firefox: not tested. On the macOS 27 machine this was tried on, neither Playwright's Firefox build nor a stock Firefox 157 would start under automation (Playwright's build hangs at start-up; the stock build exits with "Could not find profile folder"), so no Firefox call has run. The likely route is Playwright Firefox on Linux CI; until that or the manual checklist has passed, Firefox stays "not yet verified".
Manual check (Safari and Firefox)
This takes about 15 minutes per browser. You need two devices, because one device cannot play both people properly (see "One camera per browser" above). Use your Mac plus your iPhone, or your Mac plus a second computer.
Run it once for each browser you want to tick off:
- A. Mac: Safari (the "first device") with iPhone: Safari (the "second device").
- B. Mac: Firefox (first device) with iPhone: Safari (second device).
Steps:
- First device: open https://demo.synento.com. If the page says "Browser not supported", stop and send a screenshot. That is a failure.
- Click Create new session. A long code appears in "Session ID". Type a
name in "User ID" (for example
mac), then click Join Live Session. Allow the camera and microphone when asked. Pass: "Status" says Connected and you see yourself. - Second device: open https://demo.synento.com, paste the same code
into "Session ID" (send it to yourself by Messages or AirDrop), type a
different "User ID" (for example
phone), tap Join Live Session, and allow camera and microphone. Pass: each device shows two pictures, itself and the other, and the other person's picture is moving, not black. - Speak into each device in turn (keep them in different rooms, or turn the volume down, to avoid feedback). Pass: you hear each side on the other.
- Type a message in the chat box ("Say something…") and click Send, on each device. Pass: the message shows on both.
- iPhone only: switch to another app for about 15 seconds, then come back. Pass: within a few seconds the status returns to Connected and the Mac sees the phone's picture again. A short gap is expected.
- Click Leave on both devices.
- On the Mac, sign in at https://app.synento.com with the account that owns the demo, open Recordings, find the session by its code, and click Replay. Pass: both people play back smoothly, and you hear both of them, in step with their lips. If the button next to the speeds reads Enable sound, click it once first. Browsers hold sound back until you click.
- In the replay, click Download MP4, then open the file. Pass: it plays, with picture and sound.
How to report back: for each step, write the step number and pass or fail, which browser and device it was on (for example "Mac Safari 27" or "iPhone 17, iOS 27"), and the session code. For any failure, add a screenshot and what you saw instead. If a picture stays black or there is no sound, say on which device.
iOS
Two behaviours are worth knowing about, and both are platform behaviour rather than Synento bugs:
- Backgrounding stops capture. iOS suspends media capture when the tab is not frontmost. The connection resumes on return, but the recording has a gap.
- Autoplay of remote audio needs a gesture. Audio playback starts from a user interaction (the join tap counts). A page that connects on load without any tap may render video silently until the user touches something.
What your code should do
checkSupport() returns a structured report, and SynentoRoom.connect() refuses
to open a socket when it fails, rather than half-joining a call:
import { SynentoRoom } from '@synento/client'
const support = SynentoRoom.checkSupport()
if (!support.supported) {
// support.reason is written for humans and names the missing capability.
showUnsupportedMessage(support.reason)
return
}The individual booleans (webSocket, webCodecsVideo, mediaDevices,
audioWorklet) are all on the report, so you can tell "update your browser"
apart from "this page needs HTTPS".