Launching today
Ansight
Recursive self improvement for mobile apps.
13 followers
Recursive self improvement for mobile apps.
13 followers
Ansight gives your coding agent the ability to operate mobile apps and inspect what happens inside them. It brings UI interactions, logs, network requests, crashes, performance and app state together in one queryable session. Through a CLI, your existing agent can reproduce bugs, investigate causes and verify fixes on simulators, emulators and real devices, with recorded evidence you can review. Built for native and cross-platform apps, with custom tools for your app’s own logic.
Hey everyone, I’m Martin, co-founder of Ansight.
Coding agents can write a fix and tell you the tests pass, but that doesn’t always mean the app works. A screen can show “saved successfully” while nothing reaches the database. You still end up opening the app, reproducing the issue and figuring out what actually happened.
We built Ansight to give coding agents access to the running app. They can interact with it and inspect its UI, logs, network requests and runtime state, so they can investigate the problem and check whether their fix worked.
It works with the coding agent you already use. The idea is to spend less time taking screenshots, copying logs and explaining what went wrong, and give the agent the evidence to work through it.
Would love to hear how you’re checking your agent’s changes today. What do you still find yourself testing manually?