SCAPTION is coming soon — sign up to be notified when it’s ready.

NDI-native live captioning · 24 languages · nova-3

Every word. Every language. Live.

SCAPTION turns live speech into translated captions in under two seconds — delivered as a transparent NDI source or chroma-key HDMI, straight into the AV workflow you already run.

No burn-in, no browser overlays, no post-production. Your captions arrive as a clean alpha layer your switcher composites like any broadcast graphic.

3 free hours · valid 14 days · no card · full features

SCAPTION Studio — Plenary Room ON AIR
LangESEN128 msStableDante Rx 1–2
TRANSCRIPT · CO-PILOTnova-3

Bienvenidos. Gracias por acompañarnos hoy.

Creamos herramientas para quien dirige eventos en directo.

La latencia media se mantiene bajo los 128 milisegundos —

incluso con público en el auditorio.

1 low-confidence word — click to correct before it reaches the screen

OUTPUT
Alpha NDISCAPTION · 192.168.1.24:5961

This is how we bring latency below 130 ms.

99.2%
Accuracy
1080p
Alpha out

Transcribe & translate across 24 languages — auto-detect the source

EnglishEspañolFrançaisDeutschItalianoPortuguêsNederlandsPolskiРусскийУкраїнськаTürkçeالعربيةहिन्दी日本語한국어中文BahasaSvenskaDanskNorskSuomiΕλληνικάČeštinaRomânăEnglishEspañolFrançaisDeutschItalianoPortuguêsNederlandsPolskiРусскийУкраїнськаTürkçeالعربيةहिन्दी日本語한국어中文BahasaSvenskaDanskNorskSuomiΕλληνικάČeštinaRomână
< 2s
Glass-to-glass caption latency
24
Languages + auto-detect source
99%+
Word accuracy on nova-3 STT
1080p
True alpha NDI output
HOW IT WORKS

From microphone to the venue screen

Three steps, one operator, no post-production.

01

Speak

Any WDM audio source feeds the console. Deepgram streaming STT and DeepL translation run through SCAPTION’s relay — provider keys never touch the venue machine.

02

Correct

The Co-Pilot shows the live transcript with low-confidence words flagged. Fix a name or term before it reaches the audience — editing never stalls the feed.

03

Composite

Captions leave as a 1080p alpha NDI source (or chroma-key HDMI). Key them onto the LED wall, switcher, or vMix exactly like any other input.

SEE IT RUNNING LIVE

Captions your switcher treats like any other input

SCAPTION outputs a true alpha-channel NDI source — no downstream chroma stage, no burn-in. Key it over the program feed exactly like a lower-third graphic.

Program feed
Así reducimos la latencia a menos de 130 ms.This is how we bring latency below 130 ms.
SCAPTION · alpha NDI
Así reducimos la latencia a menos de 130 ms.This is how we bring latency below 130 ms.
ON AIR
On the venue screen
BUILT FOR THE BOOTH

Everything the show needs — nothing it doesn’t

True alpha over NDI

A real alpha-channel NDI source that keys directly onto LED walls, switchers, and vMix — with a chroma-key HDMI fallback when there’s no NDI gear.

Operator Co-Pilot

A live, editable transcript. Low-confidence words are flagged so you can correct them before they hit the big screen. Editing never stalls the feed.

Built for the booth

A dark, vMix-like console. Pick any WDM audio source, set a source→target language pair, style captions to the venue’s brand, and export the transcript afterward.

Keys, not compromises

Deepgram STT and DeepL translation are relayed through SCAPTION’s backend — provider keys never ship in the app, and usage is metered server-side, not trusted from the client.

Pre-flight safety

SCAPTION warns when Windows’ Public network profile would hide your NDI source, and always shows its own name + ip:port for manual entry in vMix.

Never cut off mid-event

Hitting a plan limit warns and auto-bills overage — it never interrupts a live session. A dropped connection fades the output to transparent, never an error card.

WHO IT’S FOR

Built for live events

Rental houses, venues, worship tech teams, and corporate AV.

Conferences

Multilingual keynotes and breakout rooms. Caption the stage in the speaker’s language and translate for the room, live.

Houses of worship

Translate the service for a multilingual congregation and key the text over your ProPresenter / NDI wall.

Corporate & town halls

Accessible captions for all-hands and hybrid events, composited into your existing vMix or TriCaster show.

Broadcast & streaming

A real alpha NDI source your switcher treats like any other input — no downstream chroma stage.

WHEN IT MATTERS

Live is unforgiving. So is SCAPTION.

The event happens once. The tool has to behave when it counts.

Runs on your machine

The caption engine and NDI output are local. The internet carries audio and text, not your show — so the venue feed doesn’t depend on a browser tab staying open.

Clean, never frozen

If the connection drops, the audience output fades to transparent — never a stale caption or an error card left on the big screen.

Never cut off mid-event

Hitting a plan limit warns and auto-bills overage; it never interrupts a live session. Entitlement is only re-checked between events.

Priced per event, not per seat

Start free, then move to a plan that fits how often you go live. Overage auto-bills so a session is never cut mid-event — and entitlement is only checked between events.

FAQ

Frequently asked questions

What exactly comes out of SCAPTION?
A 1080p NDI source with true alpha transparency — your switcher composites it like any lower-third graphic. No NDI gear? Use the chroma-key HDMI fallback instead. There is no burn-in and no browser overlay.
How fast are the captions?
Speech reaches the venue screen as translated captions in under two seconds, glass-to-glass. The operator can correct any word before it is shown, without stalling the feed.
Do I have to bring my own Deepgram / DeepL keys?
No. Speech-to-text and translation are relayed through SCAPTION’s backend, so the provider keys never ship inside the app and usage is metered server-side. Everything is included in your plan.
What does the free trial include?
Three hours of full-featured captioning and translation, valid for 14 days, no card required — enough to prove it on a real event before you pay.

Run your next event free

Three hours of full-featured captioning and translation, valid 14 days, no card — enough to prove it on a real event before you pay.

Start free trialDownload for Windows