Sports organisation
A sports organisation streaming lower-tier or youth fixtures that currently go out without any commentary
AI & automationReady to pilotBuilt by LabsLast updated 29 July 2026
Sports CommentatorLive match commentary generated from the video feed itself
Lower-tier fixtures go out silent because a human commentary crew cannot be justified for the audience they draw. Sports Commentator generates live play-by-play narration that reads a real event feed, tracks match momentum, and shifts its tone from calm build-up to a goal reaction without a person in the loop.
Real-Time AI CommentaryContext-Aware NarrationMulti-Provider TTSEmotion-Driven Delivery
The problem this solves
For every top-flight fixture with a commentary team, there are dozens that go out with nothing but crowd noise. The economics are straightforward: a crew costs more than the advertising revenue a small audience can generate, so the fixture is deprioritised or streamed silent. Viewers accept it because there is no alternative, but it leaves a material gap in the product for organisations with tiered competition structures.
Pre-recorded summaries and synthetic narration from static scripts have been attempted before, but both fail on the same point: they do not react. A goal is read in the same flat tone as a substitution. Commentary that cannot respond to what is happening on the pitch does not feel like commentary at all.
Who this is for
This is relevant if you are:
Broadcaster
A broadcaster with a large fixture slate who cannot scale a human commentary operation to cover every match
Rights holder
A rights holder looking to increase the value of second- and third-tier distribution rights
How it works
- 1Live event data is ingested from the Sportradar API: passes, shots, fouls, cards, corners, substitutions and score changes, each stamped with a match clock timestamp.
- 2An emotion and intensity engine reads the match state - pitch position, score context, upcoming events via lookahead - and sets a tone level from calm through normal, tense and excited to intense, which also drives the pacing of the delivery.
- 3A large language model writes the commentary line for each event from the tone level, the tactical context and the running score, so the output is a sentence somebody could have said rather than a template with the names swapped in.
- 4Each line goes to a TTS engine - ElevenLabs or OpenAI - with its own emotion tags and delivery instructions, which is what keeps a shouted goal call and a quiet passage of play from sounding the same.
- 5A synchronisation layer aligns each line to its match clock timestamp, detects the silences and fills them with contextual material - summaries, atmosphere, background - so the audio track never drops out. That alignment is only as good as the timestamps in the feed, which is where the offsets described under Current status come from.
What makes it different
Most automated sports audio is either scripted - which means it cannot respond to the game - or generated from a flat prompt that treats a goal and a throw-in identically. This system evaluates match state continuously and adjusts both what it says and how it says it. The tone engine is the thing that makes the output sound like commentary rather than a running log, and it is what distinguishes this from a text-to-speech layer on top of a statistics feed.
Current status & next steps
Ready to pilotReviewed 29 July 2026
Sports Commentator is ready to pilot. The system has been run against a recorded fixture with live-style event data, which represents the current state of the build. The recording is not published here for now - ask us and we will walk you through it.
Synchronisation is not yet exact. In many passages the narration lines up with the action, but commentary can also arrive slightly early or slightly late relative to the footage, because the event data from our provider (Sportradar API) is not always precisely time-synchronised - offsets of a few seconds are possible.
The next step is a pilot against a live fixture calendar, so timing accuracy and tone calibration can be measured on real match conditions.
Interested in bringing commentary to every fixture?
Tell us which tier of competition you are looking to cover and roughly how many fixtures per week, and we will tell you what a pilot would involve.