Features

From the encoder to the player, in one install.

Ingest, adapt, record, and play — written the way an operator would explain it, not as four vendors stapled together.

Contribution

Ingest

Publish from the encoder, the room, the file, or a URL already on the network.

WebRTC / WHIP: Browser publish, and WHIP from encoders that already speak it.
SIP audio/video: Room systems and phones that join over SIP, with audio and video.
RTMP: From OBS, vMix, Wirecast, and hardware encoders on the desk you already have.
RTSP: Cameras and recorders you already installed, pulled or pushed into the same server.
SRT: Contribution across a long or lossy link, without pretending the network is perfect.
VOD: A file you already have, served from the same install as the live path.
From URL: Pull a stream that is already published at an address you can reach.
NDI: Switchers, graphics, and cameras already on the studio network, received without a capture card in the middle.

Playback

WebRTC / WHEP: Sub-second play in the browser, including players that speak WHEP.
HLS: Playback playlists a CDN can cache, for people who are only watching.
RTSP: Hand the stream to a player or recorder that still expects RTSP.

Distribution

Adaptation

Codec conversion: Automatically transcodes audio and video codecs when they are not supported between protocols.

Coming in a future release

Inside the path

01
Transmuxing
Change protocol or container without a re-encode when the formats allow it.
02
Recording
MP4 and HLS files, plus a window viewers can rewind while the session is still live.
03
Thumbnails
Poster frames and sprite sheets so a player can scrub without guessing.
04
Audio
Level, transcode, and mix per output when the room has more than one source.
05
Markers
Cue points and captions that travel with the picture, for the app that displays them.

Day to day

Sizing: One install, sized for the sessions you run. Docker, a binary, or a package.
Observability: Prometheus metrics, structured logs, and a trace you can follow per stream.
Where it runs: A datacenter, a cloud VM, or a network that never touches the public internet.
A second machine: Put your own load balancer in front and keep a warm standby. Vasl does not invent a cluster for you.
Upgrades: Drain viewers, then restart. Shutdown hooks exist so you are not cutting a live room by surprise.

Access

Token auth: Short-lived JWT keys for publish and play. When the appointment ends, the key does too.
Signed URLs: Links that expire, and that you can pin to an address if you need to.
SSO and roles: SAML or OIDC for the people who operate the server, with roles that match the job.
Encryption: TLS on the wire, SRTP for WebRTC, AES-128 for HLS.
REST and webhooks: Create streams, rotate keys, and hear when a recording is finished.
SDKs: Web, iOS, Android, and a server-side client.

Try it on your own traffic