A Mac app that sits in your menu bar, sees every AI tool you use — Claude Code, Cursor, Copilot, all of them — and holds the context that makes each session smarter than the last. Your preferences, your project history, your corrections — all compounding quietly in the background.
A simple Mac app. Sits in your menu bar. Learns while you work.
Works with the AI tools you already use
No setup. No prompts to write. No AI expertise required. Stash observes, learns, and gradually gives you superpowers you didn't have to build.
Which projects are open. Which AI tools are running — Claude Code in one terminal, Cursor in another, Copilot in a third. On day one, it also finds your existing setup: project instructions, agent configs, connected services. You start with momentum, not a blank slate.
When you tell your AI "no, do it like this" — Stash captures it. After a few similar corrections, it creates a reusable skill. But it goes deeper than memory: Stash builds a living picture of your work — your preferences, your standards, your project context, your history — so every AI session starts with everything it needs instead of nothing.
Skills apply automatically across every project and tool. Recurring tasks become one-click flows you can delegate. An advisor watches each project so you can ask "what happened?" instead of digging through logs. Work that started as collaboration gradually becomes delegation — and you didn't configure anything to get there.
Not another tool to learn. A quiet layer that makes every tool you already use smarter.
At first, you work alongside AI — shaping, correcting, iterating together. Over time, Stash captures enough patterns to let you hand off entire workflows. That progression happens naturally.
Building an app together. Debugging a tricky issue. Iterating on a strategy doc. The AI is your collaborator — but it doesn't know your preferences, your project history, or what you tried yesterday. Every session is a blank slate. The AI is powerful, but the context is missing.
A weekly blog post. A recurring data pull. A standard onboarding checklist. The work has enough structure and trust that you can hand it off as a flow — AI follows the steps, applies your skills, and surfaces the result for your review. You went from doing it to directing it.
Stash lives in your menu bar and automatically detects every AI tool running on your Mac — no matter who made it. See every project, every AI session across every tool, what needs your attention. Send tasks to any project without switching windows. Queue work for later. Your AI keeps moving while you stay focused.
Every project gets an advisor that sees everything: what AI is working on, what files changed, what skills are active, what you said your goals were. Ask it "what just happened?" or "draft the next task" and it answers with full context — not a generic chatbot, an advisor that's been watching your specific project. It speaks when asked. Never interrupts.
You don't create skills. You don't configure anything. You just work. Stash notices when you keep correcting the same thing and offers to remember it permanently — whether you're in Claude Code, Cursor, or Copilot. One skill applies everywhere. Your preferences compound silently in the background. Your AI improves while you sleep.
You've told your AI to "always check customer reviews before recommending a vendor" three times this month — across two different projects. Want to make this permanent?
Yes, remember thisHave a task you do every week? Tell Stash about it and it turns it into a reusable flow. But here's the magic: Stash also watches for patterns. Do the same kind of work three times and Stash offers to capture it automatically. Either way, next time you hit "Run," fill in the details, and AI handles it with the right instructions, context, and skills attached.
Which AI sessions finished overnight. Which ones need you. What changed across all your projects. When work in one place affects another, Stash catches it and passes context along — so nothing falls through the cracks, even when you're not watching.
Every project progresses through five levels — automatically, just from you doing your normal work. What starts as hands-on collaboration gradually becomes trusted delegation.
AI tools are powerful. But without context, every session starts from scratch. Stash holds the context — and makes it compound.
Each workstream gets an AI advisor with full context — your goals, what the agent did, which skills are active. Ask "what happened?" or "draft the next task" and get answers that know your project, not generic responses.
Corrections become skills. Repeated work becomes one-click flows. And Stash imports what you already have — project instructions, agent configs, connected services — so you start with leverage, not from scratch.
Claude Code, Cursor, Copilot, Codex, Cowork, Gemini — Stash detects and coordinates them all. No AI provider will ever build a tool that monitors their competitors. Stash is the only layer that sees everything.
When changes in one project affect another, Stash catches it. It passes context between AI sessions automatically and warns you before agents step on each other. No per-project tool can do this — they only see their own world.
Most AI memory just accumulates. Stash self-improves: skills get tested against real output, weak ones get flagged, strong ones get applied more broadly. The context doesn't just grow — it gets sharper. Month 6 is dramatically better than month 1.
You're already running Claude Code, Cursor, or Copilot across multiple projects. You don't need help prompting. You need one surface to monitor everything, an advisor that actually knows your project context, a task queue that keeps agents moving, and a way to see what all of it costs you.
You switch contexts all day. You come back to a project and spend 20 minutes remembering where you left off. You want one place that shows everything — and you want AI that picks up where it stopped, with full context, without you re-explaining.
Stash is the on-ramp. Start by working alongside AI the way you already do — Stash watches, captures patterns, and builds skills in the background. When it notices something you do repeatedly, it offers to turn it into a flow you can delegate. You go from "using AI" to "directing AI" without studying anything.
Install a Mac app. Work the way you always do. Stash learns your patterns, builds your skills, and gradually turns collaboration into delegation — across every project, every tool, everything you touch.