Find the moment.
Check the answer.
Verificate Hybrid-VLM is Compile-Time Inference applied to video. Broadcast footage from News24:7 is time-stamped and indexed as it airs, so a question about what was shown comes back in plain English with a link to the supporting clip. You check the answer by watching.
External demonstration on public broadcast footage. Use public, non-sensitive questions. Availability may vary.
From broadcast to a clip you can check.
Every frame stamped as it airs
The base layer is the News24:7 live broadcast feed. Each frame is time-stamped and indexed the moment it airs — a running, cited record of what was shown and when.
Your archive writes on top
A newsroom or rights-holder layers its own footage, transcripts and questions over the live base. The combined store is yours.
Answers link to the clip
Ask a question and the answer comes back in plain English with a link to the supporting clip. Check the claim by watching it.
Cite, or decline
Where the footage supports an answer, it links to that moment. Where it does not, the intended behaviour is to say so rather than fill the gap. The language model describes the selected footage.
A claim about what aired should be checkable by watching, not taken on a chatbot's word. Two things are worth keeping apart: a signed, time-stamped recording tells you where footage came from and when it aired. It does not, by itself, prove that what was said on screen is true, or that an interpretation of it is correct. This is a demonstration — talk to us about your feed, archive and questions.