Introducing EventRecast
A real-time captioning platform for live events — built on the bet that within a few years, every live event will produce a caption stream the way it produces a video stream.
Today we're shipping the first public version of EventRecast.
EventRecast is a real-time captioning platform for live events. Broadcasters publish live captions to attendees on any device. Translation runs in the same pipeline so international audiences read in their preferred language. After the event ends, an AI summary, a searchable transcript, and engagement analytics are immediately available.
This post is about why we built it, what we learned from talking to the event organizers and accessibility coordinators we built it with, and what's in the platform on day one.
The gap that made us build this
Captioning a single keynote used to require a stenographer in the room. Captioning every breakout, panel, and Q&A across a multi-track conference was financially out of reach. The math simply didn't work — at hundreds of dollars per hour per room, a 4-track conference over 2 days cost more than the rest of the AV budget combined.
So most events captioned the keynote and skipped everything else. Accessibility coordinators wrote that gap into program documents that weren't supposed to exist. Deaf and hard-of-hearing attendees got served by the keynote and abandoned by the rest of the schedule. International attendees fell back on translation apps that lagged the speakers by ten seconds. Anyone in a noisy environment — half the conference floor — had no fallback at all.
Modern speech recognition and the modern browser changed what's possible. The cost curve flipped: captioning a sixth, seventh, twelfth concurrent room costs nearly nothing more than captioning the first. That's the threshold above which 'we caption every session' becomes a real operational stance instead of a marketing claim.
Three principles we ship around
Captions belong on the audience's device, not the speaker's screen. Per-device captioning lets each attendee control font size, contrast, and language. Accessibility decisions live with the attendee, not enforced for the room. This also keeps the in-room and remote experiences consistent for hybrid events — both audiences open the same viewer URL.
Latency is the only metric that matters. Captions that lag the speaker by more than four seconds stop being read. We tune the entire pipeline — capture, transcribe, distribute — for sub-three-second end-to-end latency. Most events come in under two.
The transcript is owned media, not a byproduct. Every captioned event becomes a searchable, indexable, shareable artifact that lives at a stable URL. That's where most of the long-tail value comes from — months after the live event, the transcript keeps producing attention, content, and search visibility.
What's in the first version
Real-time captioning to any attendee device, with sub-three-second target latency. Custom vocabulary per event so technical jargon and product names land correctly on the first try.
Multi-language translation, viewer-side. Each attendee picks their language from the viewer page. One broadcaster reaches an English audience, a Spanish audience, a Mandarin audience simultaneously.
AI summaries at the moment the event ends. Executive summary, key takeaways, action items, topic tags — pulled from the actual transcript, not fabricated.
Searchable transcripts with speaker labels, timestamped lines, and shareable deep-links to specific moments. Exportable as SRT, VTT, JSON, or plain text.
Engagement analytics: concurrent viewers over time, drop-off curves, language splits, most-shared moments per event. The data sponsor reports actually need.
Who EventRecast is for today
Conference organizers running multi-track programs. University and corporate L&D teams running recurring lectures. Accessibility coordinators operationalizing WCAG 2.2 / ADA / Section 508 captioning programs. Conference platforms embedding captioning into their products without building the captioning pipeline themselves.
If any of those describe you, the homepage links to a use-case page that goes deeper on how it fits. Or skip ahead and start a free trial — the marginal cost of trying it on your next event is roughly zero.
What's next
We're building the next-quarter roadmap with the customers running EventRecast on real events. The patterns we keep hearing: deeper LMS / event-platform integrations, better support for live editor workflows, an embed SDK for event platforms, and the third or fourth tier of post-event content automation.
If your event program would benefit from any of those, we'd like to hear about it. Live captioning is a piece of infrastructure that quietly enables a lot of downstream work; the better we understand the downstream, the better the platform gets.