Meta opens up Muse AI gadget code and ships a limited Home Link device

Meta has released code that lets people build their own Muse-powered gadgets and has started offering a limited-run Home Link device that can connect the company’s new AI agent to lights, TVs, printers, and other home gear. The move, announced on October 2, matters because it pushes Meta’s Muse beyond a software demo and into a more practical, do-it-yourself hardware ecosystem.

The company’s pitch is straightforward: if you have an off-the-shelf ESP32 board or a Raspberry Pi, you can use Meta’s newly opened code and SDKs to attach Muse to displays, buttons, sensors, actuators, and other components on your workbench. Meta is positioning the idea as an open-ended hardware playground for developers and tinkerers, while also warning that users are on their own when it comes to safety and reliability.

At the same time, Meta is giving away a prebuilt companion device called Muse Home Link. According to Nat Friedman of Meta Superintelligence Labs, the company manufactured 5,000 units and is letting interested users join a waitlist ahead of shipping later this month. The Home Link is intended to work with community-built skills that let Muse carry out household tasks based on a user’s setup.

What Meta just released

Meta has taken a significant step toward making Muse more than a standalone AI agent by publishing code for hardware makers and hobbyists. The company says people can now connect Muse to their own displays or tiny embedded systems and build custom devices around it.

That opens the door to a range of homegrown gadgets: an AI reminder panel on a color E Ink screen, a living-room display running on an HDMI stick, or a compact touchscreen device that resembles a custom version of Meta’s own Muse Charm product. In Meta’s framing, Muse is no longer just something you interact with in a controlled app or prototype environment. It can become the brain inside a physical device.

How does Meta want people to build Muse gadgets?

Meta says builders can use an ESP32 development board or a Raspberry Pi, then wire Muse into components such as displays, buttons, sensors, actuators, and any other hardware they have available. The company’s language suggests a broad maker-friendly approach rather than a tightly controlled consumer platform.

In practical terms, that means Muse can be embedded in projects that monitor a room, trigger actions, or surface AI responses on a screen. The company is encouraging experimentation, but it is not promising polished consumer behavior out of the box.

Meta describes Muse gadgets as open-source devices that users build themselves and connect to their own hardware, while cautioning that they should proceed at their own risk.

Why this matters for Meta’s AI strategy

This release hints at a broader strategy: Meta appears to be trying to make Muse useful in physical spaces, not just in apps and chats. By opening the code, the company is inviting developers to treat Muse like a platform layer that can sit inside everyday objects.

That approach matters for two reasons. First, it creates a path for enthusiasts to build novel AI gadgets without waiting for Meta to ship every possible product variation. Second, it helps Meta test how people may want to use an AI agent in the home, which is a much messier and more diverse environment than a smartphone screen.

It also reflects a familiar Meta tactic: seed a platform early, encourage community experimentation, and let outside builders expand the ecosystem. If that works, Meta can gain momentum, feedback, and product ideas at relatively low cost.

What makes Muse different from a normal smart display?

Muse is meant to be more than a static dashboard. Meta is presenting it as an AI agent that can react to connected hardware and community-built skills, meaning it can potentially do more than show information or respond to basic commands.

That distinction is important. A smart display is mostly a screen. A Muse gadget, in Meta’s vision, is a programmable device that can interact with the physical world through sensors and actuators, which expands the range of possible use cases.

Muse Home Link: Meta’s own hardware bridge

Alongside the open-source release, Meta is also distributing Muse Home Link, a device the company built itself to help connect Muse to home equipment. Meta says the gadget can use community-created skills to carry out tasks such as switching on a light, controlling a television, sending a file to a printer, and other actions depending on the user’s setup.

The Home Link sounds less like a finished consumer appliance and more like a bridge between Muse and the smart devices already sitting in people’s homes. In effect, it gives Meta a ready-made reference device for showing how Muse can operate in a real household.

Nat Friedman, now associated with Meta Superintelligence Labs, said the company produced 5,000 units. Users can sign up for a waitlist now, with shipping expected sometime this month.

How the open-source approach changes the picture

Meta is not merely offering a preview or a closed demo. By opening the code, it is handing builders the tools to adapt Muse to whatever physical form they want. That means the ecosystem can evolve in ways Meta itself may not have anticipated.

Open hardware-style launches often work by creating a community of tinkerers who are willing to test early ideas, document their discoveries, and improve the software stack over time. For a product like Muse, that can be especially valuable because the most useful applications may be discovered through experimentation rather than top-down product planning.

What kinds of projects could appear first?

The most likely early projects are simple, visible, and easy to demonstrate. Those include reminder boards, information panels, status displays, and home-control widgets. Builders are likely to favor projects that make Muse’s behavior obvious to others, especially if they want to show off the combination of AI and hardware.

More ambitious examples could follow, particularly if hobbyists and developers begin linking Muse to cameras, motion sensors, or other connected components. Even so, Meta’s own warning suggests the company is aware that experimentation may outpace safety and usability.

What Meta is not promising

Meta’s announcement leaves plenty unresolved. The company is not presenting Muse gadgets as finished consumer appliances with guaranteed reliability. Instead, it is encouraging people to build their own versions and emphasizing that they do so at their own risk.

That matters because hardware projects can fail in ways software-only products do not. A device that controls lights, printers, or televisions has to deal with power, networking, device compatibility, and user expectations. In a home setting, mistakes can be frustrating at best and inconvenient at worst.

Meta’s caution also suggests that the company is prioritizing exploration over polish. The release looks designed to accelerate learning, not to present a fully mature smart-home platform.

How does this fit into the broader AI hardware race?

Meta’s move fits a larger trend in the AI industry: companies are searching for ways to make agents useful outside the browser and the app drawer. As AI systems become more capable, the next battleground is increasingly about where they live and how people physically interact with them.

Some companies are focusing on wearables, others on assistants built into phones or computers, and others on dedicated devices. Meta’s approach is unusual because it is leaning into a maker ecosystem rather than a tightly controlled mass-market product category. That could be a strength if developers rally around it, or a weakness if the tooling remains too experimental for everyday users.

For Meta, there is also a branding dimension. The company has long been interested in hardware, from smart displays to mixed-reality products. Muse gives it another entry point into the connected-home conversation, this time with the added appeal of AI agents and open-source development.

Timeline of the Muse rollout

The following timeline captures the key steps in Meta’s latest Muse hardware push.

• October 2: Meta announces that it has open sourced code for custom Muse gadgets.
• October 2: The company confirms it has created the Muse Home Link device for home control use cases.
• October 2: Meta says 5,000 Home Link units were produced.
• October 2: Interested users are directed to a waitlist.
• This month: Meta expects the Home Link devices to begin shipping.

What developers and hobbyists should know

Anyone interested in building with Muse should understand that this is an early-stage ecosystem. Meta is providing code and suggesting possible hardware combinations, but the company is not promising a plug-and-play consumer experience.

That makes the announcement especially relevant for developers, makers, and smart-home enthusiasts who enjoy assembling their own systems. The use of accessible hardware such as Raspberry Pi boards and ESP32 controllers lowers the barrier to entry, which could help the idea spread quickly among hobbyists.

Still, the practical value of the release will depend on whether the community builds compelling skills and whether the SDKs are stable enough to support them. The Home Link device may help by giving users a reference point and a shortcut for getting started.

Why the limited hardware run is notable

Producing 5,000 Home Link devices is not a mass-market launch, but it is enough to seed a small user base and gather feedback. That number suggests Meta is testing the waters rather than betting the company on immediate consumer adoption.

Limited runs like this often serve several purposes at once. They build excitement, create scarcity, and allow a company to monitor real-world use before scaling. For Meta, that seems especially important because the product category is broad, experimental, and highly dependent on developer adoption.

What comes next?

The key question now is whether builders will actually embrace Muse as a hardware platform. If they do, Meta could end up with a small but lively ecosystem of AI gadgets that range from practical home tools to playful prototypes.

If they do not, Muse could remain a curiosity tied to Meta’s own demonstrations and limited distribution. The difference will likely come down to whether the code is easy to use, the hardware integrations are reliable, and community-built skills are useful enough to justify the effort.

For now, Meta has done the most important thing: it has turned Muse from a company-owned demo into something others can build with. That is often the first step in turning an AI product into a platform.