Why We’re Building a Linux Computer You Can Wear
Date: July 26, 2026 Speaker: Lyle, Founder of Monako
We recently hosted Monako’s first community AMA, where we discussed our product vision, the move to Qualcomm AR1, our open Linux environment, spatial computing, voice interaction, the Monako Ring, and the progress of our first production batch.
The most important question was:
｜What exactly are we trying to build with Monako? Monako is not simply another pair of smart glasses connected to an AI assistant. Our goal is to build a genuine wearable Linux computer—an open computing device designed for developers, AI agents, and the next generation of human-computer interaction. Here are the key takeaways from our first AMA.👇
1️⃣·Moving to Qualcomm AR1
Monako was originally based on a Rockchip platform. As the project evolved, we realized that a device focused mainly on displaying text could not support our broader vision. We therefore decided to make a major hardware upgrade: the first generation of Monako will now be built on Qualcomm AR1. AR1 includes an NPU delivering approximately 2.4 TOPS, allowing us to explore more on-device computer-vision and spatial-perception workloads, including: • Hand tracking • YOLO-based object detection • Head-pose tracking • IMU sensor fusion • Visual-inertial odometry • On-device spatial visual processing Changing platforms required parts of our drivers, software stack, and emulator to be adapted or rebuilt. This is one of the main reasons our developer tools were delayed. The transition added complexity, but it also significantly raised the technical ceiling of what Monako can become.
2️⃣·An Open Linux Computer, Not Another Closed Device
Most consumer devices use an app-centric model in which applications operate inside restricted environments. That model works well for traditional mobile software, but AI coding agents naturally depend on access to: • File systems and project directories • Shell commands and local processes • Package managers • Network requests • Persistent read-and-write workspaces Monako is taking a different path. Developers will be able to connect through SSH, access the file system, execute system commands, and install their preferred development environments. Monako is designed to support: Lua · C · Python · JavaScript · Node.js · Bun Our long-term goal is simple: ｜If a program can run on Linux, we want to make it possible to run that program on Monako. Developers should not have to wait for us to build a dedicated API for every possible use case, nor should their applications be confined to a closed app-store ecosystem.
3️⃣·An SDK Designed for Developers and AI Agents
Monako will provide an SDK, but in a more open form than a traditional mobile SDK. Our planned developer resources include: • Open-source example applications • A Lua SDK • Display and sensor examples • Voice-agent examples • System images • Backup and recovery tools • Agent-readable code and documentation These resources will explain the device not only to human developers, but also to AI coding agents. A developer could give Monako’s example projects to Claude Code, Codex, or another coding agent. After reading the code, the agent could learn how to render an interface, access microphones, process sensor data, and make network requests—then generate a new application from a natural-language request. Our official examples will primarily use Lua, with: • Skia for 2D interface rendering • libuv for networking and asynchronous tasks • C for low-level hardware integration • Rust and WGPU for spatial rendering Lua is our recommended starting point, not a restriction.
4️⃣·From a Monocular Display to Spatial Computing The first generation of Monako uses a monocular green waveguide display. During development, we discovered that head tracking and spatial anchoring can create a workspace much larger than the physical display—even with a monocular system. Our early prototype: 1.Renders application interfaces with Skia
2.Converts those interfaces into textures
3.Maps them onto 3D surfaces
4.Anchors windows around the user using head-pose data
A developer could place a terminal to the left, keep the main task in front, and position additional agent windows to the right. Turning their head would allow them to move naturally between these workspaces. We are currently developing: • AHRS-based head tracking • 3DoF spatial windows • Hand tracking • YOLO-based object recognition • VIO-based camera positioning • A longer-term path toward 6DoF
Some capabilities already exist as prototypes, while others remain in active development and optimization.
5️⃣.Why Spatial Displays Matter for AI Agents We believe more work will become agent-driven. People may increasingly describe an intended outcome—writing code, researching a subject, creating a presentation, or managing a workflow—and let agents perform much of the execution. But humans will still need an essential interface: ｜A screen on which to observe, verify, and manage the agent’s work. A phone is too small, while a laptop limits users to the physical screen they can carry. Glasses can potentially create a virtual workspace much larger than the device itself. A user could view: • Multiple coding agents • Terminals for different projects • Current task status • Agent-generated results • Calendars, messages, and reminders • Information related to the surrounding environment Monako is therefore not merely an input device for AI. It could become a portable display, terminal, and Linux computer for the age of AI agents. It will support both independent on-glasses computing and connections to more powerful remote computers or servers.
6️⃣·Voice, Microphones, and Physical Interaction Monako will include a foundational voice-agent application that users can use immediately and developers can fully replace. Developers will be able to: • Replace the ASR or TTS provider • Connect a different AI model • Modify prompts and workflows • Connect a private server • Rebuild the voice experience entirely We have tested services including FunASR and Cartesia, while continuing to evaluate local quantized speech models. Following community feedback, we decided to keep the bone-conduction microphone. The current design includes: • 1 bone-conduction microphone • 3 conventional microphones This system is intended to support quiet voice commands in noisy environments, beamforming, meeting transcription, face-to-face translation, and environmental audio processing. We also continue to believe in tactile physical controls. Monako includes a small scroll wheel that can be rotated and pressed, providing more reliable feedback than an invisible touch surface.
7️⃣·The Monako Ring The Monako Ring is optional and will not be required to use the glasses. Users will still be able to operate Monako through: • Physical controls • Voice interaction • Visual gestures The Ring is intended to extend these interactions through: • Health tracking • Accelerometer-based motion sensing • Touch and swipe input • Waking the glasses • Motion-based spatial controls By combining touch and motion sensors, the Ring could eventually function as a miniature spatial controller. The Ring is not designed for NFC payments. Its focus is health sensing and human-computer interaction.
8️⃣·First-Generation Hardware Targets Based on our current engineering targets, the first generation of Monako includes: • Qualcomm AR1 • Approximately 2.4 TOPS NPU • Open Linux operating system • Monocular green waveguide display • Approximately 48 g weight • 1 bone-conduction + 3 conventional microphones • Open-ear speakers • Physical camera cover • Wi-Fi and Bluetooth • Prescription-lens support • Physical scroll wheel and buttons • Confirmed first-batch price: $399 Planned prescription options include: • Single-vision lenses • Progressive lenses • 1.61 refractive index • 1.74 refractive index Some battery, storage, software, and production specifications are still undergoing final engineering validation. We will publish confirmed figures once that process is complete. For transparency, extended monocular-display use in a completely dark room is not recommended, and the first generation will not include eye tracking.
9️⃣·Freedom, Security, and Ownership An open Linux system gives users more freedom—but also greater responsibility. Because developers can access the complete file system, an incorrect script could damage the operating environment. Open systems also require greater care when handling credentials and private information. Our plans include: • Complete downloadable system images • Self-service system recovery • Backups and snapshots • Hardware-backed storage for sensitive credentials • A physical camera cover We will not pretend that an open system is free of risk. Our philosophy is not to decide what users should be allowed to do with their own device. It is to give users genuine ownership while providing the recovery and security tools that ownership requires.
🔟·Building the First Batch with Our Community Moving to Qualcomm AR1 significantly increased both development costs and the bill of materials. Despite this, we will honor the confirmed $399 first-batch price. Based on current estimates, the first batch may generate little or no margin, and we may absorb a meaningful loss on each device. We are willing to do this because our first users represent more than orders—they are early collaborators helping us build the product. Community feedback has already influenced major decisions, including: • Keeping the bone-conduction microphone • Opening the complete Linux environment • Prioritizing agent development • Adding more reliable physical controls • Providing recovery tools and system images • Exploring Ring-based interaction Production is moving through engineering, design, and production validation. We will continue sharing real progress instead of promising dates before the underlying engineering is sufficiently certain.
🚀 What Comes Next We plan to release the following progressively: • Lua SDK • Open-source example applications • Skia interface development environment • Spatial-window demonstrations • Voice-agent examples • Developer documentation • Staged emulator releases • On-device vision demonstrations Our priority is to help developers start building as early as possible while ensuring that the tools we release are genuinely usable. Monako is still at an early stage. Many capabilities remain under development, and many engineering challenges still need to be solved. That is exactly why we have chosen to build in public. We do not want to decide alone what smart glasses should become. We want to define the next generation of personal computing together with our community. Welcome to Monako.👓