A Computer Systems Plus company · Fresno, CaliforniaSystem statusHosted in the USAsales@civicca.com

Sending meeting audio to live captions over RTMP or SRT

SentraCaptions team · · 3 min read

Most agencies already stream their meetings from an encoder such as Pearl, OBS or vMix. This guide explains how to send that same stream to SentraCaptions over RTMP or SRT, so you can add live captions without new hardware.

The idea: one more destination

Your encoder probably already sends your meeting somewhere, such as YouTube Live, your meeting video platform, or both. SentraCaptions works the same way. You add it as one more destination on the encoder you already use, and it listens to the meeting audio to produce captions and translations.

That means:

RTMP or SRT: which to use

SentraCaptions accepts both. Each meeting gives you the connection details you need for either protocol.

RTMPS (RTMP over TLS)

RTMP is the protocol most encoders and streaming platforms have supported for years. If you already stream to YouTube Live, you are likely using RTMP today. It is usually the quickest to set up because the fields look familiar: a server URL and a stream key. RTMPS is the same thing with TLS encryption, the protection HTTPS uses, so the meeting audio is encrypted on its way to us. Use the rtmps:// server address when your encoder supports it; most current versions of OBS, vMix and Pearl do.

Plain RTMP (the rtmp:// address) also works for older encoders, but the audio is not encrypted in transit.

SRT

SRT is a newer protocol designed to hold up better on less reliable networks. If your encoder supports SRT and your IT team prefers it, it is a good choice. Setup is similar: you enter an SRT address with the details provided for the meeting. On a permanent channel with SRT encryption turned on, the address includes a passphrase and the audio is encrypted with AES-256.

If you're unsure, start with RTMPS. You can switch later without changing anything else about the meeting.

Step by step

The exact menus differ by encoder, but the steps are the same.

  1. Create the meeting in SentraCaptions. Give it a title and choose the languages for that meeting (up to 8 from the 75 supported).
  2. Copy the stream details. Copy the RTMPS server and stream key, or the SRT address, for that meeting.
  3. Add a destination on your encoder. In Pearl, OBS, vMix or your encoder of choice, add a new stream output or channel destination and paste in the details. Keep your existing destinations as they are.
  4. Start streaming. When the encoder starts streaming to SentraCaptions, captions begin to appear about one second after speech. The sentence in progress updates live as the speaker continues.
  5. Check the chamber screen and phone page. Open the chamber screen on the TV in the chamber or lobby and scan the QR code with a phone to confirm captions and translations are showing.

Test before the meeting

A short test before your first live meeting saves stress later. A simple routine:

Common questions from IT

Does this change our public stream?

No. SentraCaptions is an additional destination. Your existing outputs continue as before. If you also want captions on YouTube Live, captions in one language can be sent to the YouTube Live stream. See Adding live captions to your YouTube Live meeting stream.

What does SentraCaptions need from the stream?

Captions come from the audio. Clear audio from the chamber microphones matters far more than video settings. See Chamber audio tips for better captions.

Can we run more than one room?

Yes. Several rooms or meetings can run at once, each with its own stream details, its own chamber screen and its own link for residents.

What about recess and closed session?

Staff can pause captions. While paused, nothing is captioned, translated, stored or billed. When the meeting resumes, captions pick up again. See Pausing captions for recess and closed session.

Can our platform start meetings automatically?

Yes. The SentraCaptions API lets a platform create a meeting, stream to it, read captions as JSON or Server-Sent Events, and receive a webhook when it ends.

If you use CivicCA, captioning starts when the meeting starts and pauses for recess and closed session, so clerks don't need to manage it separately during the meeting.

After the meeting

When the stream ends, transcripts are available in every language you selected, as TXT or SRT (subtitle) files. SRT files can be added to your posted meeting recording. Transcripts are not official minutes, but they can help staff prepare them.

Learn more