Multimodal by default
Type, dictate, paste a screenshot or share a video walkthrough. The platform reads each input as a first-class instruction, so you can move at the speed of thought instead of the speed of forms.
What's supported
Inputs
Voice
Push-to-talk dictation with on-the-fly correction.
Screenshots
Drop a screen and ask 'why does this convert badly?'
Video walkthroughs
Upload a Loom or screen recording — @iRen scrubs the frames.
PDF / spec docs
Drop a brand book or a brief, the agent extracts and applies.