Launching today

Dialect
Give the physical world an API
5 followers
Give the physical world an API
5 followers
Point any camera at something you'd otherwise walk over to check: a bay, a shelf, a valve, a door. Say what to watch in plain English. Dialect's agent sets up a typed endpoint with nothing to train, keeps reading it, and serves the answer over REST, MCP and webhooks, with alerts. When it isn't sure, it says so instead of guessing. Your first answer is free; $99 a month covers your first thing.






Payment Required
Launch Team

AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted
Hey Product Hunt, I'm Tyler, and I built Dialect.
Software stops at the physical world. Is the bay free? Did the valve open? Is the dog in his crate? Someone still has to walk over, look, and type it in.
Dialect turns that look into an API. You point a camera at the thing and say what you want to know. An agent sets up the endpoint, tests it on what the camera sees, and starts reading. There's no model to train and nothing to label. Your software reads it over REST, your agents over MCP, and webhooks tell you the moment it changes.
The thing I care about most: when Dialect can't tell, it says it's unsure rather than guessing. On our held-out test cameras it's wrong under 2% of the times it says it's sure.
Your first answer is free, so you can see whether it handles your question before paying. I'd love to hear what you'd point it at.
@tyler_gibbs What are some of your best use cases? I think this has lots of potential.