Point OBS at a URL and let people watch your stream in a browser
One service, one configuration field, no code, and you have an RTMP endpoint and an adaptive HLS stream. That is the whole thing.
What you get
An adaptive HLS stream with two renditions, produced by the service defaults. These are the exact renditions we measured by pushing a test pattern to the RTMP endpoint and decoding a frame back out of the live manifest.
| Resolution | Video codec |
|---|---|
| 1280x720 | avc1.42c01f |
| 640x360 | avc1.42c01e |
Audio is AAC-LC. The output framerate follows your input, so the service imposes no frame rate of its own.
How to do it
- 1
Create the service instance
Create an instance of Eyevinn Live Encoding. The only required configuration field is a name. No access token is needed and no code is written.
- 2
Wait for the certificate
A new instance reports ready before its TLS certificate has been issued, so the first request can fail. Wait about a minute and retry the same request unchanged.
- 3
Point your encoder at the RTMP endpoint
Set the RTMP endpoint of the instance as the server in OBS Studio, or in any other RTMP encoder, and start streaming.
- 4
Serve the HLS manifest to viewers
The adaptive HLS manifest is served at /origin/hls/index.m3u8 on the instance host. Give that URL to any HLS player in the browser.
What it costs, and which plan you need
This stack runs at 50 tokens per day. It is heavy for a single service because it carries an encoder that runs continuously, where an ordinary service costs 10 per day.
The free Basic plan grants 100 tokens one time, not per day, so it runs this stack for under two days. Personal gives 40 tokens per day, which is below what this stack needs to stay up. Professional is the plan that runs this continuously.
We are telling you this before you deploy rather than after, because the tiers are sized for ordinary services and this is not one. See pricing for current plans.
For agents and tooling
The machine-readable descriptor for this stack is served as JSON, so an agent can select it without parsing this page. It names the job, the service, the required inputs, the produced outputs, the full configuration surface and the cost.
/stacks/live-rtmp-to-hls.jsonBuilt entirely on unmodified open source. The service is Eyevinn Live Encoding, which is ffmpeg based, and you can run it yourself anywhere instead if you prefer.