AI Assistant for Work · Local & Private

The AI assistant that already knows your work.

Every other AI assistant expects you to copy-paste context, re-upload documents, and explain your week from scratch. Maple is an AI productivity assistant that runs on your computer, understands your active screen and meetings locally, and delivers answers grounded in your real work context.

Get early access Download for Mac & Windows

Runs locally · Private on-device indexing · Zero data training

AI ASSISTANT

The trouble with generic web chatbots

General-purpose chatbots know the public web, but they know nothing about your actual workday.

The Copy-Paste Tax

To get a useful answer from ChatGPT or Claude, you spend five minutes scrubbing and pasting Slack threads, Jira tickets, and meeting summaries into a prompt box.

No Persistent Memory

Every browser session begins as a blank slate. Your AI personal assistant forgets the technical decisions you made yesterday and the sprint goals established this morning.

Cloud Privacy Anxiety

Pasting proprietary financial models, client strategies, or codebase diffs into third-party cloud servers risks sensitive data leakage and policy violations.

How Maple operates as your AI assistant for work

A native desktop companion engineered around contextual awareness, privacy, and calm computing.

Screen-Aware AI

Maple observes your active window title and screen content on-device. When you tap ⌥ Space, it answers with direct reference to what you're currently editing. Learn about screen-aware AI →

AI Memory Assistant

An on-device SQLite vector index connects past conversations, notes, and documents into a persistent knowledge graph that never forgets. Explore Maple's AI memory →

AI Meeting Assistant

Captures live calls on Zoom, Google Meet, and Teams via driver-level audio loopback—transcribing and extracting action items without inviting awkward bots. See meeting assistant features →

Local-First Security

A private AI assistant built for confidentiality

Maple is developed with a strict local-first architecture. Raw screen frames, microphone audio, and document indexes remain stored exclusively on your physical hardware. When you ask a query, only the specific retrieved context snippet is dispatched to frontier AI models with strict zero-retention policies.

Read our privacy architecture

Frequently asked questions

What makes Maple different from generic AI chatbots?

Generic AI chatbots operate entirely in the browser and rely on whatever text you manually copy into the prompt. Maple is a native desktop AI assistant that indexes your screen, meetings, and local documents in real time, so answers are already grounded in your active work.

Does this AI assistant upload my screen recordings to the cloud?

No. Screen analysis and text extraction happen directly on your CPU/GPU using localized processing. Your raw screen captures are never stored in a cloud database or transmitted over the network.

Can Maple assist with team meetings and calls?

Yes. Maple includes native system-level audio capture for Zoom, Meet, and Teams. It transcribes discussions, tags speakers, and extracts deliverables locally without adding third-party recording bots to the call.

Upgrade your workday with an AI assistant that remembers.

Join our early preview cohort. Available for macOS and Windows.

Get early access