Floatboat
Proactive AI agents that run your work from your calendar
7 followers
Proactive AI agents that run your work from your calendar
7 followers
Floatboat is a proactive agent OS for calendar-driven work. Turn meetings, deadlines, routines, and plans into automated agent workflows that run at the right time. Each task lives in an Agent Workspace that keeps context, files, decisions, and agent actions together. Built for teams and multi-device work so humans and agents stay aligned without losing context.
Hey Product Hunt 👋
I'm Judy, cofounder of FloatBoat. I used to be a tech community builder, then a VC — and for years, I watched people drown in the same cycle: back-to-back meetings, zero prep time, forgotten follow-ups, and a calendar that only tells you WHERE to be, never WHAT to do.
ChatGPT is great for conversations. But it waits for you to ask. Your calendar doesn't wait — meetings just show up. So you're always scrambling: researching attendees 5 minutes before a call, forgetting action items by end of day, never sending that follow-up.
That's exactly why we built FloatBoat.
It's a proactive agent OS that lives on your desktop and runs from your calendar. Connect Google Calendar, Notion Calendar, or Lark — and AI agents take over the work around your meetings:
• Pre-meeting brief? Auto-generated with attendee backgrounds, recent news, and talking points — before you even open the invite.
• Post-meeting follow-up? Drafted with action items and next steps, ready to review and send.
• Recurring workflows? Subscribe any calendar and trigger Deep Research, slide decks, or content prep automatically via Combo Skills.
• Complex tasks? Use FloatIM — a group chat where every member is an AI agent that delivers finished work, not just text.
We built this for busy professionals who don't need another chatbot — they need an AI that watches their calendar and does the work before they ask.
Your calendar already knows what's coming. FloatBoat makes sure you're ready.
Judy & the FloatBoat team
What makes Floatboat interesting to me is that it doesn’t just live in a chat box. It brings together context, files, browsing, memory, and execution in one continuous workflow, which makes it feel much closer to an operating system for knowledge work. That’s a big difference from most AI tools, which still feel like isolated prompt-response interfaces. With Floatboat, the experience feels more connected, more persistent, and more aligned with how real work actually happens. That’s why it stands out to me as a much more interesting direction.
Seeing Floatboat in action is truly a delight. A friend of mine works at a university. When drafting speeches previously, she noticed that models could construct standard structural layouts. For example, they would allocate the first paragraph for national policies, the second for university background, and the third for thematic elevation. Lacking specific factual data support, however, the models failed to provide accurate policy texts or concrete university details, resulting in hollow content and forced thematic transitions. Consequently, she would only retain the AI-generated framework and rewrite the entire text sentence by sentence.
After Labor Day, she tried using Floatboat to assist with her writing and observed a remarkable transformation in the output. The structural integrity remained, and the specific details were flawlessly accurate. When she suspected certain phrasing deviated from the Ministry of Education's official terminology, she instructed Floatboat to retrieve relevant official press releases and formal speeches. The model immediately executed the search and replaced the text with precise official language.
Floatboat's harness layer is already optimized to an exceptional degree.
The 'Combos' feature is a game changer. Being able to bundle specific skills and reuse them across workflows is exactly what was missing from my current stack. It’s ambitious to call it an 'AI Office Environment,' but after 5 minutes, I can see why. Great job, team! One question: Do you plan to support more local IDE integrations?"