We built Maple because context is everything.
The Genesis of Maple
It started with a simple, daily frustration shared by developers, designers, and product teams: the best AI chat models make you repeat yourself. You open a new window, and the model has zero memory of what you did five minutes ago. You find yourself constantly copying and pasting files, screenshots, active errors, and slack chats just to explain a simple bug or ask for a draft response. You are serving as a manual data pipeline for your AI.
We realized this wasn't a limitation of model intelligence. Rather, it was a limitation of model sight. The model is smart, but it is blind to your workspace. So we set out to build cognitive infrastructure that runs locally on your machine, learns your professional context, and is ready to assist you the second you hit ⌥ Space (Mac) or Alt + Space (Windows).
Our North Star
We aren't building another generic chatbot. Our objective is to create the first AI system that feels like a real colleague — one that already has your context, understands your week's priorities, and respects your privacy boundaries by default.
The Four Core Pillars of Maple
To guide our engineering decisions and product roadmaps, we commit to four foundational principles:
- Privacy is Non-Negotiable: We believe your intellectual property and workspace conversations belong exclusively on your hardware. Maple uses local-first indexing. Your files, Slack chats, database credentials, and calendar events are parsed, vectorized, and cached in an encrypted SQLite database on your physical device. We never run cloud syncs, and we never train models on your data.
- Context Over Parameter Size: Knowing your codebase structure, your meeting summaries, and your design choices is far more valuable than running a trillion-parameter general model with zero knowledge of your project. By indexing your local files and workspace integrations, we supply hyper-relevant prompts to lightweight, lightning-fast inference models.
- Silent and Non-Intrusive by Design: Modern software is filled with attention-hijacking notifications, pop-up recommendations, and auto-opening sidebars. Maple respects your focus. It sits quietly in your tray or menu bar. It does not send unprompted reminders or alerts. It appears only when you call it, does its job, and disappears the moment you click away.
- Respecting Device Resources: Running background AI engines shouldn't turn your laptop into a noisy heater. Maple's context crawler is written in native Rust. It monitors your CPU temperature, system load, and battery state. If you unplug from power or run heavy compilation tasks, Maple throttles its background processes instantly.
Independent
Maple is built by a tight-knit, independent group of designers, software engineers, and AI researchers based out of Jabalpur, Madhya Pradesh, India. By choosing to remain self-funded and founder-led, we avoid the pressure of growth-at-all-costs metrics. This allows us to focus entirely on crafting a premium, high-fidelity user experience for people who care about calm computing and software craftsmanship.
We build and ship in the open. Our design decisions, roadmap priorities, and engineering challenges are shared weekly on our journal. If you're interested in shaping the future of context-aware software alongside us, explore our open batches.
Explore open batches and roles
A Calm Workspace
Maple is designed to reduce the mental overhead of daily context switching. We measure our success not by how much time you spend inside our interface, but by how quickly you can get out of it and back to your creative work.