Live translation for government meetings: how it works

Live translation for government meetings sounds complicated, but the path from a council member's microphone to a resident's phone is short. Here is what happens at each step of real-time translation, what residents see, and what staff need to do.

The path from microphone to screen

Most agencies already stream meetings through an encoder such as Pearl, OBS or vMix. Live translation rides along on that same stream.

  1. Audio leaves the chamber. You add SentraCaptions as one more destination on your existing encoder, over RTMP or SRT. The encoder keeps sending video to your usual places; it just sends a copy to us too. No new hardware.
  2. Speech becomes English text. Speech recognition turns the audio into English captions about one second after the words are spoken. The sentence in progress updates live, so readers see words appear and settle as the speaker finishes a thought.
  3. Text is translated. Machine translation turns each caption into the languages you picked for that meeting, up to 8, from a list of 75 languages and regional variants.
  4. Captions go to screens. English and translated captions appear on the chamber screen, the phone page and, for one language, your YouTube Live stream.
  5. Transcripts are saved. When the meeting ends, transcripts are available in every language as TXT or SRT files.

Note that this is real-time translation of text. Residents read the meeting in their language; they don't hear it. For spoken interpretation you still need a human interpreter. See translation vs. interpretation for the difference.

Where residents see the translation

The chamber screen

A full-screen caption display for the TV in the chamber or lobby. It can show one language or several side by side, along with the meeting title, a clock and a QR code. Setup tips are in our chamber screen setup guide.

Residents' phones

Scanning the QR code opens a phone page where residents pick their language and text size and read along. There's no app to install and no sign-in, which matters for residents who are wary of handing over personal information. It works for people watching from home as well as those in the room. More in captions on residents' phones.

YouTube Live

Captions in one language can be sent to your YouTube Live stream. YouTube shows one caption language per stream, so translations live on the chamber screen and phone page. See YouTube Live captions for meetings.

What to expect from the translation

Machine translation is fast and covers many languages, but it is not perfect. Plan around a few realities:

Running it during a meeting

Once the encoder destination is set up, day-of work is light:

If you use CivicCA, the start and pause steps are handled for you: CivicCA starts captioning when the meeting starts and pauses it for recess and closed session. Captions also show on the CivicCA dais display and public portal.

How this fits the rules

SB 707 does not itself require live captions or live translation. For city councils and boards of supervisors that are “eligible legislative bodies,” it does require translated agendas in applicable languages, reasonable assistance to members of the public who want to translate a meeting or receive interpretation, and outreach to communities that don't traditionally participate, with most of those requirements starting July 1, 2026. Live translated captions are a practical way to support those language-access and participation goals. Check with your city attorney or county counsel on what applies to your body.

Separately, the ADA Title II web rule points to WCAG 2.1 Level AA, which includes captions for live audio in synchronized media (SC 1.2.4). Read more in WCAG live captions explained.

Next steps