TVU Networks Powers VSiN at the Circa Sports Book: 19 Live Hours a Day, AI-Clipped and Published in Under Two Hours, Zero Added Editors

Downtown Las Vegas. Inside the Circa Sports Book, a five-story display towers over the room. A few steps away, in the Vegas Stats & Information Network (VSiN) control room, artificial intelligence is listening to every word. For 19 live hours a day, TVU Search’s AI transcribes the broadcast, understands what the analysts are talking about, and cuts it into on-demand clips, with no editor touching a timeline.

VSiN delivers 24/7 coverage as “the MSNBC of sports,” and it moves at that pace. Lines shift on air. The Dodgers move to +1320. The Brewers go to -10,000. Every segment is built for the bettor watching right now.

The problem is what happens to that segment an hour later. With TVU Search, VSiN found a way to make every minute of live analysis work twice: once on air, and again as an on-demand clip, published in under two hours.

The Challenge

Sports betting content has a brutal expiration date. A pick matters until the final whistle. After that, it is worth nothing to the person who wanted it.

“Our content has a short shelf life… whatever pick or whatever side we thought was going to win that game, it doesn’t matter for our sports bettor once the game is over.” — Miles Gwyn, Chief Operating Officer, VSiN

To get more value from its live output, VSiN needed to cut 19 hours of continuous broadcast into targeted digital clips, fast. The traditional way means loggers, editors, and copywriters watching every feed, marking timecodes, trimming segments, writing metadata, and uploading to OTT and social.

VSiN ran the numbers. Doing that across its full daily output would take a dedicated team of 8 to 10 editors. For a lean network, that was not an option.

The Solution: TVU Search

VSiN met TVU Networks at the NAB Show and deployed an AI-driven clipping pipeline built on TVU Search. It ingests the live feed, listens, and packages clips in real time.

TVU Search does not wait for a human to spot a cut point. It transcribes the broadcast audio as it airs and reads the meaning of the conversation. When the desk moves from Monday Night Football to late-night West Coast college basketball, the AI hears the shift and slices the block into single-subject clips. Each one is indexed with structured metadata and routed to the right content stream.

“We’re able now to say this one segment where we talked about these three games is not a segment going over three games—it’s these three games that are now their own individual clips.” — Miles Gwyn, Chief Operating Officer, VSiN

Off-the-shelf AI tends to stumble on betting talk: odds formats, line movement, granular player evaluations. So VSiN and TVU trained the speech models on VSiN’s own lexicon, down to a debate over whether a right tackle “may have been the worst player on the field.” The result is automated clips that meet broadcast standards.

The Results

VSiN now turns on-air commentary into published clips in under two hours, without adding a single editor. That speed opened doors a lean control room could not reach before:

• Pop-up channels. VSiN launched specialized feeds such as College Football Best Bets powered by VSiN with no extra production staff.
• A self-updating app. VSiN rebuilt its mobile app around video playlists that fill themselves by sport, matchup, and marquee event.
• Easier syndication. Every clip carries structured metadata, so VSiN can offer partners live, API-ready feeds instead of packaged file transfers. Integration friction drops, and new licensing deals open up.

“We always like to say internally we can move at the speed of thought… this workflow is an example of exactly that.” — Socrates Lozano, VP, Solutions, TVU Networks 

Nineteen hours of fast-expiring linear TV now feeds a multi-channel digital operation serving millions of viewers.

Why It Matters

VSiN is an early example of where live media operations are heading. For years, turning a live feed into on-demand assets meant hiring a room full of people to watch it. AI that listens, understands domain language, and writes its own metadata removes that bottleneck.

The lesson is not just automation. It is automation trained on the business. VSiN’s clips are accurate because the models learned how bettors and analysts talk.

Any newsroom where content loses value by the hour faces the same choice, from financial news to breaking live events. VSiN made it early. Its content still expires when the game ends, but now it earns its keep before it does.