I'm Seni, solo builder of Ash. I started building this because I was drowning in my own email, tasks, and context-switching and every AI assistant I tried either forgot everything between conversations, or wanted my inbox and files sent to someone else's cloud to work at all. Neither felt right for something meant to actually manage my life. So I built Ash to run entirely on my own computer — a local LLM, memory that persists across days and weeks, and real follow-through on tasks instead of just answering when asked. Nothing about how I use it daily gets sent anywhere. My approach changed a lot along the way. I assumed the hard part would be features —more integrations, more tools. It wasn't. The hard part turned out to be reliability: making sure she actually does what she says instead of just sounding like she will. Most of my recent work has gone into finding and closing exactly that kind of gap, not adding new capabilities. Still early and rough in places genuinely want your honest feedback, especially if you've thought about the privacy/trust tradeoffs of giving an AI deep access to your inbox and files. Happy to answer anything about how it works under the hood too.

Ash — AI Chief of Staff

A local-first AI chief-of-staff for your inbox and tasks

A local-first AI chief-of-staff for your inbox and tasks