Two things at once. A feed of real video from the places around you, and an assistant that has watched all of it, and can answer what no general chatbot can.
Check out the real app, filmed on real streets.
Not a scroll of anywhere. Every clip is tied to the spot it was shot, so a neighbourhood becomes something you can actually watch. It's why people open the app, and it builds a record of the city that exists nowhere else.
Ask is a concierge for the city you're standing in. It reads that record (the footage, the audio, the words on screen) alongside live events and the open web, so its answers come from what actually happened, not from a guess.
Neither is a feature of the other. The feed is what makes Ask worth asking; Ask is what makes the feed worth more than a scroll.
Pulse is the city: everything being filmed around you, anywhere in the world. Friends is the people you follow. Shoot vertical, post, and it lands on the map where you stood.
Posting takes seconds. Titles and descriptions are written from what's actually in the clip, so you film and move on.
The whole city, or just your people. Filter by city or go global.
Vertical video from the app. The place comes with it.
Clips from one outing group themselves, with a written arc.
No hardcoded home town. The app works the same wherever you open it.
Ask runs on Contrivance's own multimodal models. They've watched the video, heard the audio and read the text on screen. Ask about the street you're on, and the answer comes with the clips it came from.
Where's good tonight. What was that festival last weekend. And on any clip: what's that jacket, where is this, what did he just say, answered from the video itself.
Ask Home opens with what changed in your city overnight, and what it drew on.
Answers cite the clips and sources behind them.
Ask any clip a question. It sees the frames, not just the caption.
Type it, say it, or attach a photo.
Events, venues and the clips filmed around them, on one map. Filter to what's free, what's music, what's on tonight, then build a plan and ask about the area without leaving it.
The map is the city's calendar and its camera roll at once: what's on, where, and what it actually looks like.
Live events with real listings, filtered by what you care about.
Put a question to the patch of map you're looking at.
Save what you're going to, and get it back as a plan.
Talk to the people behind the clips. Direct messages, and voice calls that are end-to-end encrypted.
A night out becomes one story, written from the clips in it rather than assembled by hand.
Built multi-city from the start. Open it in a new place and it works the same way.
Cohlta is made by Contrivance Inc., a New York company building multimodal AI: software that understands video, audio and text together. The models behind Ask run on hardware we own, not on someone else's API.