An RTMP streaming server without having to run Nginx
We give your encoder an RTMP URL and key, and we take care of the adaptive bitrate transcoding, the HLS output, the recording, the restream to YouTube/Facebook and the CDN. 24/7 human support. Full setup in 5 minutes.
Encoder compatibility
The RTMP server receives video from any tool or device that sends RTMP/RTMPS, which is practically the whole industry:
| Category | Supported tools |
|---|---|
| Free software | OBS Studio, Streamlabs, Larix Broadcaster (mobile), GoCoder |
| Pro software | vMix, Wirecast, XSplit, ManyCam |
| Hardware encoders | BlackMagic ATEM Mini Pro, Teradek Vidiu/Bond, AVMatrix, LiveU Solo, Epiphan Pearl |
| IP/PTZ cameras | BirdDog, Datavideo PTC, NewTek Spark |
| Mobile | Larix Broadcaster (iOS/Android), Streamlabs Mobile |
How to configure OBS with our RTMP server in 60 seconds
- Create your account. TestDrive+ plan free for 3 days, no card.
- Open OBS Studio. Menu Settings → Stream.
- Service: select "Custom".
- Server: paste the RTMP server URL from the panel (e.g.
rtmps://ingesta.xtreamcast.co/live). - Stream key: paste the stream key from the panel.
- Apply and then click Start Streaming. The public HLS URL goes live immediately.
If you want a visual guide with screenshots: how to stream professionally with OBS (in Spanish).
RTMP vs RTMPS: secure ingest to your RTMP streaming server
When you configure your encoder against our RTMP streaming server, you choose between two flavors of the same protocol. Plain RTMP travels unencrypted over TCP on port 1935, the historical standard inherited from Adobe. RTMPS is exactly the same RTMP but wrapped in a TLS layer, just like HTTPS does with HTTP; it travels encrypted over port 443, the same as the web.
| Aspect | RTMP | RTMPS |
|---|---|---|
| Encryption | No (plaintext over TCP) | Yes (TLS, same as HTTPS) |
| Typical port | 1935 | 443 |
| URL prefix | rtmp:// | rtmps:// |
| Passes corporate firewalls | Sometimes (1935 is often blocked) | Almost always (443 is open) |
| Latency impact | Baseline | Negligible (minimal TLS overhead) |
Recommendation: use RTMPS whenever your encoder supports it —OBS, vMix, Wirecast and virtually all modern gear do—. It encrypts the broadcast between your encoder and our ingest, and by traveling over port 443 it passes firewalls that block 1935, a common scenario when streaming from corporate networks, hotels or universities. It only makes sense to fall back to plain RTMP when an old hardware encoder doesn't implement TLS. In both cases you don't need to open inbound ports: your encoder initiates the outbound connection to our live streaming servers, so allowing outbound traffic on 1935 or 443 is enough.
Per-encoder configuration for your RTMP streaming server
Four parameters matter: the server URL, the stream key, the keyframe interval and the bitrate. For stable HLS, keep the keyframe interval at 2 seconds (closed GOP): it's what allows clean HLS segmentation. This table summarizes the recommended configuration for the most common encoders:
| Encoder | Server URL | Stream key | Keyframe | Recommended bitrate |
|---|---|---|---|---|
| OBS Studio | "Custom" service + rtmps://…/live | Stream key from the panel | 2 s | 3.5 Mbps (720p) / 6 Mbps (1080p) |
| vMix | "Custom RTMP" destination + panel URL | Stream name/key from the panel | 2 s | 6 Mbps (1080p30) / 9 Mbps (1080p60) |
| Wirecast | "RTMP" output + panel URL | Stream key from the panel | 2 s | 4.5–6 Mbps (1080p) |
| Streamlabs | "Custom" service + panel URL | Stream key from the panel | 2 s | 3.5 Mbps (720p) / 6 Mbps (1080p) |
| Larix Broadcaster (mobile) | New connection rtmps://…/live | Appended at the end of the URL | 2 s | 2.5–4 Mbps (720p on 4G/5G) |
| Hardware encoder (Teradek / ATEM Mini Pro) | "RTMP" platform + panel URL | Stream key from the panel | 2 s | 6 Mbps (1080p) / per uplink |
Your upload speed should be at least double the chosen bitrate to leave headroom for peaks. If you stream over 4G/5G with an unstable uplink, lower the bitrate or consider SRT as backup. Want more detail on specs and ABR resolutions? It's in the streaming server pillar.
From RTMP ingest to HLS out: how your feed flows
Understanding the pipeline helps you diagnose problems and choose well. When you hit "Start Streaming", your feed passes through five stages inside our RTMP streaming server:
- 1. RTMP/RTMPS ingest. Your encoder opens a persistent TCP connection and delivers a single H.264/AAC stream. This is the layer where RTMP lives; the latency it adds is 1–3 s.
- 2. ABR transcoder. From that single ingest we generate several qualities in parallel (240p, 480p, 720p, 1080p) with adaptive bitrate streaming, so each viewer gets the one their connection can handle.
- 3. HLS / LL-HLS packaging. Each quality is chopped into segments and described in a
.m3u8manifest. There is no RTMP here anymore: the output is standard HLS (6–8 s) or LL-HLS (~3 s), playable in any browser without plugins. - 4. Multi-region CDN. Segments are distributed through a network of PoPs close to the viewer. This absorbs audience peaks without saturating the origin and reduces buffering wherever your audience is.
- 5. Player. The white-label HTML5 player requests the manifest, picks the right quality and switches in real time as the network varies. The viewer only sees the video.
The key idea: RTMP is the input protocol, HLS the output one. RTMP comes in because it's universal in encoders; HLS goes out because it plays on any device. If you want to dig into the differences, read RTMP vs HLS vs WebRTC (in Spanish), or jump straight to the live streaming platform.
Managed RTMP server vs running your own Nginx-RTMP
It's the decision that defines your operation. Running your own Nginx-RTMP on a VPS receives and delivers the stream, but everything else is on you: transcoding to multiple qualities, a CDN for large audiences, recording, security, monitoring and maintenance. A managed RTMP streaming server like XtreamCast integrates that full stack with a 99.99% SLA. The honest TCO:
| Criterion | Your own Nginx-RTMP | XtreamCast managed |
|---|---|---|
| Reasonable audience | <200 concurrent viewers | No viewer cap |
| Real monthly cost | VPS $20–80 + CDN $30–300 + DevOps hours | From $25/month, all included |
| ABR transcoding | Manual with FFmpeg, eats VPS CPU | Included, dedicated CPU |
| Cloud recording | Set up S3, rotation and backups by hand | On by default |
| Restream to YouTube/Facebook | Separate module, manual configuration | Up to 10 destinations from the panel |
| Monitoring and SLA | You build it (Grafana, alerts, on-call) | 99.99% contractual SLA |
| Time to first stream | Hours or days | ~5 minutes |
Rule of thumb: below ~200 concurrent viewers, and if your team enjoys operating infrastructure, your own Nginx-RTMP can make sense for learning or because of an internal network or compliance constraint. Above that threshold, the operating cost, the outage risk and your team's time justify a managed server. To decide with data, we have three deep dives (in Spanish): professional RTMP server vs self-hosted, the 2026 guide to RTMP streaming servers in Latin America and the tutorial on how to set up your own RTMP server. When you've made up your mind, compare plans and pricing.
Frequently asked questions about the RTMP server
What port does the RTMP server use, and which one does RTMPS use?
Plain RTMP uses TCP port 1935, the industry standard since Adobe. RTMPS travels encrypted over TCP port 443, the same as HTTPS, which helps it pass corporate firewalls that only allow web traffic out. You don't need to open inbound ports: your encoder initiates the outbound connection, so allowing outbound traffic on 1935 or 443 is enough.
RTMP vs RTMPS, which one should I use?
Use RTMPS whenever your encoder supports it —practically all modern ones do—, because it encrypts the broadcast between your encoder and our server and passes firewalls better. The latency difference is negligible. Only fall back to plain RTMP if your old encoder doesn't support TLS.
What is the RTMP ingest latency?
RTMP ingest adds very little latency, 1 to 3 seconds, thanks to its persistent TCP connection. The total latency the viewer perceives depends on the output: standard HLS 6–8 s, LL-HLS ~3 s and WebRTC sub-second. The bottleneck isn't RTMP, but the distribution format you choose.
Can I use my own RTMP URL or a custom domain?
The RTMP/RTMPS ingest URL is generated by XtreamCast and points to our managed infrastructure; it's the one you paste into your encoder. What you can customize with your domain is the output: the public player URL and the HLS with a white-label player. If you need to control the ingest URL end to end, that means running your own Nginx-RTMP.
Resources on streaming protocols
Ready to operate with broadcast quality?
TV channels, production companies, churches, radio stations and institutions around the world already run on XtreamCast. 3-day free trial. 7-day money-back guarantee. Your channel ready in minutes. 24/7 human support.