Guide · Events and stages
Live translation for a speech or presentation: translated captions on the audience's phones, by QR code
By Blink Brief · Updated · How pages are made
For one speech in one language, put a QR code on the screen before you start: the audience scans it with the phone's camera, picks a language, and reads each line translated on their own phone a moment after you say it, with no app. Blink Brief's free Live Captions does this from a laptop or an Android phone near the speaker, for up to 200 phones, free during the beta. It suits a keynote, a graduation address, a toast or a eulogy; for a speech people need to hear in their language, book an interpreter.
Short answer
- Before: start Live Captions, press Share to phones, and show the QR code on the screen.
- During: one speaker at a time, in one language, close to the microphone.
- The audience: reads text in their language, a moment after each line. No voice.
- After: press End sharing, which deletes the room's text.
Before the speech: the QR code on the screen or the program slide
- On a laptop with Chrome or Edge, or an Android phone with Chrome, open Blink Brief's free Live Captions. Safari works on some Macs and iPhones; Firefox doesn't.
- Choose the language you will speak, press Start captions and allow the microphone.
- Press Share to phones. A QR code and a 6-character room code appear.
- Put the code where everyone can see it before you begin: press Show the QR code full screen on a laptop connected to the projector, or have the host read out the room code for people to type at www.blink-brief.com/live-captions/listen.
Each share makes a new room with a new code, so make the code on the day, after the laptop is set up, and don't print it in the program. The program can say "Live captions in your language: scan the code on the screen." Give the room a minute before the speech: a line from the host such as "Scan the code to read this speech in Spanish, Tagalog or Vietnamese" gets phones out before the speaker starts.
If the venue has its own sound system, ask for a feed from the mixer into the laptop; the captions are only as good as what the laptop hears. A laptop's own microphone at the back of a hall picks up the room, not the speaker.
During: one speaker, one language, a phone at the podium or a laptop
- A laptop near the speaker, with the sound feed or a USB or lapel microphone, is the setup to plan for. It can also show the captions very large on the projector in the language being spoken.
- An Android phone at the podium, with Chrome, works for a toast or a short talk with no projector. Keep it close to the speaker's mouth, plugged in, and with the screen on.
- Speak in full sentences at your normal pace. Each line is translated once it is finished, so a sentence that trails off arrives late or not at all.
- Spell nothing out loud. Put names, places and numbers that matter on a slide, so nobody depends on the translation for them.
Translation is by machine (Cloudflare's AI), line by line. Read the speech through it beforehand with a speaker of each language you expect, and check names and any wording that matters.
What the audience sees (text, a moment after each line)
Each person reads in the phone's browser, on iPhone or Android, with no app or sign-up. They pick one of 32 languages and regional variants, and each line appears in it a moment after it is said; someone who picks the language being spoken reads it untranslated, which helps those who are hard of hearing. The phone page lets each reader make the text larger or smaller and switch between a dark and a light background.
- Text, not voice. Nobody hears a translation; they read it.
- A late arrival catches up: a phone that joins late gets up to the last 50 lines in the spoken language, or the last 5 translated.
- Limits: up to 200 phones and 3 hours a room, 8 languages at a time, and 100 rooms a day per network.
- Internet: every phone needs a connection, so put the Wi-Fi name and password beside the code.
- Privacy: the caption text goes to Blink Brief's server and to Cloudflare's AI for translation; End sharing deletes it, and a room nobody ends is deleted 24 hours after its last line. Your browser may send the audio to its maker's speech service (Google for Chrome, Microsoft for Edge, Apple for Safari); Blink Brief never receives your audio (Privacy Policy).
Several speakers, or a speaker who switches language
A room hears one spoken language. Several speakers in that language can share one room, one after another, as long as each speaks into the microphone the laptop hears.
- A speaker in another language: press End sharing, change the spoken language, and press Share to phones again. That is a new room, with a new QR code to show; the host can ask people to scan again.
- A speaker who moves between two languages mid-speech gets poor captions in the second one. Ask them to keep to one language, or split the speech into two parts with a new room for each.
- Questions from the floor: only what the microphone hears is captioned. Hand the questioner a microphone, or have the speaker repeat the question.
- Anything personal, such as a family matter in a toast or a eulogy: press Pause sharing first, and Resume sharing after. The phones say the speaker paused sharing.
After: End sharing
Press End sharing when the speech is over. The phones say the captions have ended, and the room's text is deleted at once. A room nobody ends is deleted 24 hours after its last line, but ending it yourself keeps the next event's audience from reading this one.
For a whole conference with sessions and rooms, see live captioning for events.
When you need an interpreter instead
Translated captions on phones help people who read. They don't replace an interpreter, and they aren't the right service when:
- Someone has asked for CART or a sign language interpreter. ADA.gov describes CART as "a service similar to court reporting in which a transcriber types what is being said". New York State's office for deaf and hard of hearing people says: "When CART services are requested, auto captions or artificial intelligence-generated captions should never be used."
- Listeners need to hear the speech, because they don't read comfortably or are elderly or have low vision. Book a simultaneous interpreter with headsets or a listening app.
- The audience will answer back in another language: a Q&A, a town hall, a negotiation. Captions go one way; an interpreter goes both.
- Every word must be right: a legal, medical or contractual statement. Use a qualified interpreter.
Interpreters, headsets and listening apps are compared in simultaneous interpretation apps.
Questions
Is there a free app for live translation of a speech?
The audience needs no app. Blink Brief's free Live Captions runs in Chrome or Edge on the speaker's laptop: press Start captions, then Share to phones, and show the QR code. People scan it with the phone's camera, pick their language, and read each line a moment after it is said, in the phone's browser. Free during the beta, up to 200 phones and 3 hours a room.
Can the audience hear the translation?
No. Phones show text, not a voice. For listeners who need to hear the speech in their language, book an interpreter with headsets or a listening app.
Can I translate a graduation speech into Spanish and Tagalog at once?
Yes. Each phone picks its own language, so one family reads Spanish and another Tagalog (Filipino) from the same QR code. A room translates into up to 8 languages at a time. Translation is by machine: have a speaker of each language read the speech through it beforehand, and check names.
Sources
- ADA.gov: ADA Requirements: Effective Communication: last updated February 28, 2020. Read on 9 November 2026.
- New York State Deaf, DeafBlind, and Hard of Hearing Office: Best practices for working with communication access. Read on 9 November 2026.