Load your script
Import a WebVTT file from your captioner. Speaker labels like RACHEL: or 1ST WITCH: work. That file is what the audience will read, every night.
Watch us make magic
Your script, in time.
Software that listens to a live show and puts up the right caption from your script. Built for D/deaf and hard of hearing access in theatres and other scripted events.
ASR transcript
Waiting for speech…
Decision
Current caption
Waiting…
Live theatre captioning is hard.
Most runs only caption a few nights. Most shows are not accessible.
-
01
A human captioner owns the words
A skilled captioner prepares the file: wording, speakers, what the audience actually needs. Bad input means bad captions. Watch us make magic shows that file; it does not invent a better one.
-
02
Then someone has to run every show
Usual practice puts an operator in the room each night, hitting next in time with the performers. That work is skilled and expensive, so venues caption fewer performances than they want to.
-
03
Spend the effort once, then caption the run
Get the caption file right, then tune the matcher from a rehearsal recording. After that you can caption every performance with the same approved text. Lighting or sound can ride the app from the desk; you do not need a dedicated operator each night.
How it works
From WebVTT to the screens in the building.
Connect your audio
Feed in a stage mic, desk mix, or induction loop. Any input device will do. Cleaner audio means cleaner matching.
Start listening
Speech recognition runs on the laptop and works out where you are in the script. It can advance, hold, catch up, or flag that the company has gone off script. Lighting or sound can keep an eye on it and override if needed; it does not need its own operator.
Send captions out
Push the current line to a projector, phones on the venue network, seat-back or lobby screens, OSC (e.g. QLab), or a webhook.
What the desk gets
The software follows the script. Lighting or sound can ride it; override stays one keypress away.
Script following with Laya
Laya is the decision model used that tracks the performance line by line.
It holds on ad-libs instead of guessing. If lines are skipped, it catches up at a pace people can still read.
Cues like [APPLAUSE] or *BANG* can fire after the line they belong to.
- Waits for evidence before moving on, so it does not race ahead
- Picks up again when the company returns to the script
- Shows your spelling and names, not a live transcript’s guesses
Desk control
Status shows ON SCRIPT, UNCERTAIN, or OFF SCRIPT. The decision panel shows which lines are in play and which words matched. Step with ← / → or Prev / Next if you need to. Manual captions (raffle, interval notices) go out without moving script position.
Audio in real rooms
Works with induction loops and mic’d performers. The speech gate follows room noise; quiet voices can be boosted up to 20 dB. Pick a lighter or larger speech model, and add custom vocabulary for names and odd spellings.
Tune from a rehearsal recording
Tuning does not write the caption file. You load your WebVTT and a rehearsal recording; the app transcribes the audio and tries matcher settings for that show. It prefers not to jump early. After a performance you can replay the session log (recognition, decisions, overrides) against different settings.
- Needs your caption file and a rehearsal take
- Searches settings; early jumps are penalised
- Replay a logged show before the next house
Where people use it
Scripted live work, mainly.
- TheatreStage screens, seat-backs, or patrons’ phones.
- Talks and panelsWhen you already have a transcript or script.
- TouringTune for the show, then run it each night.
- Opera, musicals, verseWhere the exact words matter.
- CeremoniesAwards, graduations, AGMs, set readings.
- EducationLectures with a script or speaker notes.
- Several screens at onceHouse, lobby, stream, and phones from one laptop.
Nothing leaves the building.
It runs offline. No cloud, no per-minute billing, no show audio leaving the laptop. Matching decisions are typically 50–120 ms.
- Platform macOS desktop, Apple Silicon
- Script WebVTT with optional speakers
- Models Whisper + Laya, stored locally
- Network Only for an initial model download
Where captions can go
Several outputs.
Full-screen window
Built-in display for a house screen or confidence monitor.
Local network page
Any phone, tablet, or screen on the venue LAN. No app to install.
OSC
UDP to QLab, lighting, and other show control.
Webhooks
JSON or plain text, with custom headers if you need them.
Compared with other approaches
| Approach | Difference here |
|---|---|
| Live speech-to-text | Shows your script, with correct names and spelling |
| Manual caption operator | Follows the lines itself; no dedicated outputter needed |
| Cloud captioning | Offline on the laptop; no per-minute cloud bill |
| Timecode playback | Tracks what is actually being said, including skips and pauses |
FAQ
What does it run on?
macOS on Apple Silicon. Audio from whatever input you give it: USB interface, mixer, loop receiver, and so on.
What script format?
WebVTT (.vtt), with optional speaker labels. Prepare it once, use it for the whole run.
Which languages?
English (UK/US) is what we have tested most. The speech models also cover French, German, Spanish, Italian, Portuguese, Dutch, Japanese, Korean and Chinese, but treat those as available rather than equally proven.
Can I tune from a rehearsal recording?
Yes. You supply the caption file and a recording. Tuning finds matcher settings; it does not write captions. You can also replay a logged show against new settings later.
How much does it cost?
Subscription per venue for commercial use. Charities and Arts Council England or Wales funded venues (or equivalent cultural funders elsewhere) can apply for a free license. We ask free-license users for feedback from real shows.
Who gets a free license?
Registered charities, and cultural venues funded by Arts Council England, Arts Council of Wales, or a similar public cultural body in another country. If you are unsure, email us and we will say yes or no.
Can you help set up at our venue?
Yes. Audio feed, getting captions onto your screens or LAN displays, and tuning against your file. Say whether you want someone on site or remote when you contact us.
Free for charities and cultural venues.
If you are a registered charity, or a venue funded by Arts Council England or Wales (or the equivalent elsewhere), you can apply for a free license. We can help set up at your venue for a show: audio, displays, and a tune against your caption file. We will ask you for notes from rehearsals and performances so we know what still breaks in the room.
Apply for a free license- Who Charities and Arts Council (or equivalent) venues
- What Free license for the app
- Help Setup at your venue, on site or remote
- In return Practical feedback from real shows
- Commercial Subscription per application
Get in touch
Demo, free license, provide testing feedback. Use the form and we will reply by email.