Live, and why it is paused
Live is offline. If you came looking for it, that is the short answer, and you deserve the longer one rather than a page that quietly stops existing.
What Live was
You talked, and visuals appeared behind you as you spoke.
Not slides you built the night before and then clicked through. You would be explaining how something worked, and the diagram would assemble itself while you explained it, following what you were actually saying rather than a script you had committed to in advance. If you went off on a tangent, the visuals went with you.
The use we kept seeing was teaching. Someone explaining a concept to a room, with the picture forming in time with the sentence. It is a different thing from a presentation, and people who tried it tended to understand the point within about thirty seconds of talking.
Why it is paused
The honest reason is that the first version worked in a demo and did not hold up as a product.
Getting from speech to a useful visual in under a second, reliably, for an unrehearsed talk, in a noisy room, in more than one language, is a harder engineering problem than getting there once for a recording. The version we shipped was good enough to show what the idea was and not good enough for someone to depend on in front of an audience. Live is a product where an embarrassing failure happens in public, in front of the user's own audience. That raises the bar for shipping it a long way.
So we took it down rather than leave it up while it was unreliable. We would rather have no Live than a Live that fails while you are standing in front of a room.
The engine is open, today
The visual engine behind Live is open source under the MIT licence, and it is public now: github.com/timegravity/live-visuals.
Clone it, run it on your own machine, use it, change it, build something else with it. There is no account, no waiting list and no dependency on us. If we never brought Live back at all, that part would still be in your hands, which is the point of putting it there.
If you are building something in this area, we would genuinely rather you had the engine than that we sat on it while it was not in a product.
What we intend to bring back
This section is intention, not promise. We have been wrong about timelines before, and writing "coming soon" is how a team ends up quietly not mentioning something a year later.
What we want from the next version:
- Reliability first. It has to work in a room with background noise, on
ordinary hardware, without a rehearsal.
- Visuals that follow the meaning, not the keywords. The first version
could be fooled by a word that sounded like a topic.
- More than one language, properly, not English with the others treated
as an afterthought.
- A way to keep what was generated after you stop talking, so a session
leaves you with something you can send to the people who were not there.
We are not giving a date. When it is ready to be depended on, it will be back, and this post will say so.
Tell us if you want it back
There is a form on the Live page that takes one email and sends you one message when it returns. Nothing else, no newsletter.
It is also genuinely useful to us as a signal. If a lot of people ask for Live, that changes what we work on, and the form is the only way we can tell the difference between "nobody missed it" and "people missed it quietly".
In the meantime
Chat is where the rest of the work has gone, and it is the part of the product you can depend on today. We wrote about what it does if you want the tour.
We know an offline product is a bad look, and we would rather say plainly what happened than let the page go grey. The commitments in what Timegravity is include saying when something is paused, and this is us doing that.