On May 6, 2026, Hugging Face launched an App Store for its Reachy Mini desktop robot — an open-source platform hosting over 200 community-built applications, available free of charge to device owners. It is the first initiative of its kind combining affordable robotics hardware with an agentic toolkit that allows people without any programming background to build functional robotics software in a matter of minutes.
Key highlights
- Reachy Mini App Store launches with over 200 apps from more than 150 creators — most of whom had never written a line of robotics code before
- Reachy Mini Lite retails for $299, the Wireless version for $449
- Approximately 10,000 units sold to date; 3,000 in the last two weeks alone
- Platform supports GPT-5.5, Claude Opus 4.6, Gemini Live, and several other AI models
- All apps are free and open-source; building one requires only a natural-language description
An app store for robots — what it means in practice
The Hugging Face Reachy Mini App Store functions as a repository of robot behaviors, modeled in its UX on familiar mobile app stores, but built around open hardware. A user can search for an app, install it to a physical robot with a single click, or test it in a browser using a built-in 3D simulator — without owning the physical device.
Every app is forkable: a user can duplicate an existing app and ask an AI agent to modify it with a simple instruction, for example: "have it respond in French." The entire process — from describing a behavior to a working application — is designed to take minutes, not weeks, as was the case with traditional robotics programming.
The platform is integrated with Hugging Face Hub — the company's existing infrastructure for hosting AI-powered web apps. This means all applications are publicly accessible without logging into a closed proprietary ecosystem.
Who is building apps and how the agentic toolkit works
At the core of the App Store is an agent called ML Intern — a Hugging Face tool that accepts a natural-language description of a desired behavior and generates ready-to-deploy code compatible with the robot's API. The agent handles the most technically demanding steps: translating intent into hardware instructions, adapting to firmware constraints, and testing logic before deployment.
The platform is model-agnostic. Beyond ML Intern, users can build apps using external models: GPT-5.5, Claude Opus 4.6, Kimmy 2.6, Mini Max GM5, and Deep Sig V4 Pro. For real-time conversational applications, OpenAI Realtime and Gemini Live are available.
Among the first creators, Joel Cohen — a 78-year-old retired marketing executive, colorblind, with no technical background — spent two weeks assembling his Reachy Mini Lite (a task that typically takes three hours), then built an app for running CEO peer-group coaching sessions: the robot recognizes 29 participants by name, fact-checks discussions in real time, and summarizes key themes.
Clément Delangue, CEO of Hugging Face, himself built a reception app for the company's Miami office — the robot recognizes incoming employees' faces and sends notifications. According to Delangue, it took him less than two hours, whereas "it would have been impossible before" without advanced robotics expertise.
Reachy Mini: open-source desktop robot at smartphone prices
Reachy Mini is a stationary desktop robot designed by Pollen Robotics — a startup acquired by Hugging Face in April 2025. The device features a built-in camera, speaker, and microphone. It is sold in two variants: Reachy Mini Lite ($299) — a tethered version that connects via USB to an external computer — and Reachy Mini Wireless ($449) — a standalone version with an on-board Raspberry Pi CM4 and Wi-Fi.
For comparison: Boston Dynamics' Spot costs approximately $70,000; even cheaper Chinese quadrupeds start at around $1,900. Reachy Mini positions itself well below that threshold, targeting hobbyists, educators, and developers. Since its launch in July 2025, approximately 10,000 units have been sold, with 3,000 going to customers in the last two weeks alone.
Open-source strategy as a response to the data problem
Hugging Face consistently bets on code openness. Delangue argues that closed hardware systems are "almost impossible" to train agents on at scale, because they block access to the data and code that could feed models. The open nature of Reachy Mini enables AI agents to learn hardware interaction directly from public repositories.
The Reachy Mini project is part of the broader Le Robot initiative, launched by Hugging Face in 2024. Its goal is to democratize robotics by publishing open code, tutorials, and reference hardware. Today, with the App Store, that project enters a new phase — not just providing developer tools, but creating a genuine ecosystem with end users.
Why this matters
For sixty years, robotics was a discipline reserved for engineers with deep expertise in kinematics, driver programming, and firmware specifics. The Reachy Mini App Store attacks that barrier on two levels simultaneously: through affordable hardware and through an AI agent that translates natural language into robotics code.
The effect is already visible: more than 150 creators, most of whom had never programmed a robot, built over 200 working applications. This is not a proof of concept — it is a functioning ecosystem.
An important caveat is scope: Reachy Mini is a stationary desktop robot with no locomotion capability. The applications running on it today — receptionist, language tutor, F1 race commentator — are well-defined tasks in controlled environments. Extending this approach to mobile bipedal robots or industrial manipulators is a separate, significantly harder problem. However, as a demonstration of the model — affordable hardware + open AI + agentic toolkit = creator ecosystem — the initiative is compelling. If Hugging Face can sustain this momentum, it may set the standard for how the robotics industry should think about the relationship between hardware and software.
What's next
- Hugging Face has announced the App Store will be extended with a monetization option — creators will eventually be able to charge for apps via the Spaces mechanism
- Within 30 days of launch, the company plans to ship an additional 1,000 Reachy Mini units to customers
- The Le Robot model and ML Intern toolkit are publicly available on GitHub and Hugging Face Hub — enabling rapid adoption by other robot manufacturers as a developer platform
