Lumis and the Road Ahead
When we first started Lumis, the problem was clear: I was drowning in messages. Telegram, Slack, Whatspp, Gmail, every app was competing for my attention, and I felt like I was always playing catch-up. Lumis was born to tame that chaos: summarizing, replying, and helping me stay in control.
But from day one, I knew Lumis was just the beginning.
Ingest: Breaking Down Information Silos
Everything starts with connecting the fragmented spaces where your life and work happen. Lumis acts as a universal hub, ingesting data and commands across all your platforms: Google Suites (Gmail, Docs), Slack, Telegram, Discord, WhatsApp, and more. It doesn’t just stop at text—this includes images, voice messages, files, videos, and even links. Beyond messages, it learns from your social context (who you talk to most, what you discuss) and your explicit preferences: natural language commands like “Pause auto-summaries for tech messages,” priority rules like “Flag my manager’s messages first,” or historical chats you share to refine its understanding. Advanced API integrations and distributed processing ensure no detail falls through the cracks—even as data scales over time.
Beyond Faster and Easier
The next step in communication isn’t simply about making it “faster” or “easier.” It’s about making it more personal and more intelligent. Our long-term vision is to help you build a digital twin, a representation of your memory, habits, and communication style.
This isn’t about replacing you. It’s about capturing your long-term context so Lumis can truly understand you: how you talk, how you prioritize, and what matters most. Imagine an assistant that doesn’t just summarize a chat, but knows how it fits into your ongoing projects, personal commitments, and unique way of working.
From Individual to Network
As Lumis learns from your daily use, it won’t just serve you in isolation. Over time, we see Lumis evolving into part of an agent communication network, a way for your digital twin to seamlessly coordinate with others.
Instead of one person juggling a hundred apps and chats, imagine agents talking to agents: exchanging information, aligning on tasks, and freeing humans from the noise.
Privacy as a Foundation
This vision only works if you can trust it. That’s why we’ve made deliberate design choices:
- Your raw data stays on your device.
- Sensitive tasks run locally on smaller, on-device models.
- Cloud models are used only when they add value, never for your confidential data.
If Lumis is going to become your digital twin, it has to respect your boundaries more strictly than even a human assistant would. Privacy isn’t an afterthought, it’s the foundation.
Contextual Execution with Atlas
Today, many AI products make you craft prompts or paste endless context just to get something useful. That’s backward.
Lumis should already understand enough about your daily context, your conversations, recaps, and tasks, to surface what matters without you asking.
That’s where Atlas comes in. Atlas will work alongside Lumis to handle tasks in context: turning recaps into actionable to-dos, following up on reminders, and coordinating with other tools.
Instead of you asking, “What should I tell Slack?”, Lumis + Atlas will quietly prepare the right draft or execute the task, so you can focus on what really matters.
The Road Ahead
We’re still at the beginning. Right now, Lumis is helping people stay on top of their messages. But the road ahead is much bigger:
- Digital twin with long-term memory
- Agent-to-agent communication networks
- Context-driven task execution with Atlas
It’s a journey that will take time, and it depends on feedback from you, the people using Lumis every day.
I use Lumis myself, and I already can’t imagine going back. My hope is that, step by step, Lumis becomes not just a tool you open when overwhelmed, but a true companion that works with you, understands you, and gives you back your focus.
Thanks for being part of this journey. We’re building it together.
- Ethan Founder, Meland Labs