On-Device AI Captioning for Professional AV Engineer Broadcasts.

Turn a Mac, iPad, or iPhone into an offline AI captioning server (or a Windows PC through our Chrome extension) and eliminate the cloud bottleneck for mission-critical broadcasts. Sub-1000ms latency, zero internet requirement, and seamless integration into your SDI/NDI signal chain.

Request Technical Assistance Integration Video Guide

Latency Comparison Simulator

Select a translation method and tap on Simulate Input Response to visualise the difference between local and cloud-based AI translation speeds.

                         Buffer: Waiting on network...
                    

...

The Signal Flow

Designed to sit perfectly within your existing chain.

                        
                        Audio In
                        XLR / USB / Virtual Sound Card
                        
                        ConferenceCaptioning
                        Local AI NPU Processing Core
                    
                        Captions Out
                        NDI / Display / WebVTT / 608/708

Latency: Sub-1000ms

Faster than human CART. Processing directly on the GPU/NPU eliminates transfer lag.

Air-Gapped Secure

No data leaves the local machine. Fully compliant with strict corporate and government security profiles.

Heavy-Duty Reliability

Hardware-accelerated on-device AI guarantees 100% uptime regardless of venue Wi-Fi stability.

Native Compatibility

Seamless handshakes with NDI®, RTMP, SRT streams, and output formatting as standard WebVTT.

Integration Power-User

Built for the sandbox. Our captioning engine natively interfaces with high-end production suites including vMix, OBS Studio, ProPresenter, Wirecast, Tricaster, and Blackmagic Design switchers.

Utilize our Transparent Overlay and fully Custom CSS options to inject broadcast-ready lower thirds directly into your video mix, perfectly matched to the event's precise branding specs.

Run captions from the booth. Start, stop, show, and hide captions and translations from an Elgato Stream Deck or Bitfocus Companion, right alongside the rest of your show controls. No walking back to the captioning computer mid-show.

vMix API OBS WebSockets NDI® 5 Tools Elgato Stream Deck Bitfocus Companion

Built for the "What Ifs"

AV teams tell us their biggest risk with AI captioning is depending on the venue's internet. Live production is controlled chaos. What happens when the venue internet goes dark? When thousands of attendee cell phones crash the local tower? When the designated network switch dies?

With cloud AI, your accessibility tools break.
With Conference Captioning, the translations keep scrolling.

And if the captioning computer itself fails? With the Enterprise plan, a backup Mac or iPad on the local network takes over the caption feed in as little as half a second (about two seconds at the default setting), with no page refresh, and your AV lead gets an alert email. See the hot-failover setup guide →

Deploy with absolute confidence.

REQUEST TECHNICAL ASSISTANCE INTEGRATION VIDEO GUIDE

Frequently asked questions

What is the caption latency?

Under one second (sub-1000 ms) from audio in to caption on screen. Speech is processed on the local GPU/NPU, so there's no cloud round-trip adding delay.

How do I get audio into the system?

Feed it from your existing chain: XLR or USB through an audio interface, a virtual sound card, Dante, or NDI.

Does it work on Windows, or only on Apple devices?

Both. The native app runs on Mac, iPad, and iPhone and performs best there, because it uses Apple's Neural Engine for on-device AI. On Windows, or any computer with Google Chrome, the ConferenceCaptioning Chrome extension runs speech recognition on-device too, and works offline once its speech model is downloaded. Hot-failover redundancy currently requires Apple devices.

Which production tools does it work with?

vMix, OBS Studio, ProPresenter, Wirecast, TriCaster, H2R Graphics, and Blackmagic Design switchers. Use the transparent overlay or a chroma-key background for lower thirds, and style it with custom CSS or the Caption URL Generator. See the OBS setup guide.

Can I control captions from a Stream Deck or Bitfocus Companion?

Yes. AV teams can start, stop, show, and hide captions and translations on screen from an Elgato Stream Deck or Bitfocus Companion, so the operator never has to walk back to the captioning computer mid-show.

Can it run fully offline or air-gapped?

Yes. Transcription and translation run on the local machine, so the system works on an isolated network with no public internet, and no audio data leaves the device.

What happens if the captioning computer fails mid-show?

A backup Mac or iPad on the same local network takes over automatically. The caption display checks the primary every 500 milliseconds, and after the number of missed checks you choose (4 by default), captions switch to the backup in as little as half a second, or about two seconds at the default setting, with no page refresh, while an alert email goes to your AV lead. Hot failover is included in the Enterprise plan. See the hot-failover setup guide.

Can it caption and translate at the same time?

Yes. The same feed can carry live captions plus up to 5 simultaneous translations from 50+ languages, each available as its own overlay or attendee stream.