Sunday Robotics: Who is behind Memo and can you hire it? Article summary. Sunday Robotics develops a domestic robot called Memo on wheels. You cannot hire it for your event yet, not even from ROBOTio. Who founded the company, what can Memo do, and which robot should you deploy instead? Sunday Robotics is a California-based company from Mountain View that develops a domestic robot called Memo. It is not yet available for corporate events: Memo is not sold, not available in Europe, and ROBOTio does not offer it for hire. Its beta programme for around 50 households in the USA is scheduled for the end of 2026, with commercial sales to follow. If you need a robot that welcomes guests at your event, speaks and entertains, ROBOTio deploy Unitree G1 and H2 from its own fleet, always with operators. ## Who is behind Sunday Robotics Sunday (Sunday Inc.) is based in Mountain View and was founded in 2023. It was founded by Stanford roboticists Tony Zhao (CEO) and Cheng Chi (CTO). They are the authors of ALOHA, ACT, Diffusion Policy and UMI methods - approaches that have influenced how robots learn to manipulate objects. Brief timeline: - **End of 2024:** single-arm prototype that straightened shoes, - **19 November 2025:** Memo unveiled and company emerged from stealth, - **2026:** development and beta preparation continue, the company has around 70 employees, - **End of 2026:** beta programme planned for selected households in the USA. The price of Memo has not been officially announced. You can find a full overview of the brand on the page [Sunday Robotics](https://robotio.eu/brands/sunday-robotics). ## What is Memo Memo is not a classical humanoid with legs. It is a torso with two arms mounted on a wheeled column. According to the manufacturer and press reports, it measures approximately 170 cm and weighs 77 kg. Instead of fingers, it has gripper hands. The manufacturer describes it as a domestic robot. Full specifications are summarised on the [Sunday Memo](https://robotio.eu/sunday-memo) profile. Key specifications: - **Movement:** wheeled base, speed approximately 1 m/s, intentionally half the pace of human movement, - **Reach:** from floor to approximately 2.1 m vertically, 0.8 m horizontally, - **Degrees of freedom:** approximately 20, with 7 per arm, - **Battery life:** approximately 4 hours, charged to 80% in about an hour, - **Safety:** soft silicone cover with no sharp edges, software speed limits and compliant control, so the robot can be gently pushed aside by hand. ## What makes Memo different: it learns from human demonstration The key difference is in data collection. People record household tasks using their own hands with the Skill Capture Glove, which mirrors the robot's kinematics. According to the manufacturer, more than 2,000 gloves have generated approximately 10 million demonstrations across 500+ households. The foundational ACT-1 model was trained on this data. Its successor, ACT-2, is announced for 2026. In demonstrations, Memo handles tasks such as: - loading the dishwasher: a path of approximately 40 m, 33 interactions with 21 objects including glassware, - folding laundry and socks, - preparing espresso, - putting items away in their proper places. For event organisers, this is an interesting indicator of the industry's direction. The robot is not programmed move by move, but learns from human work. However, this does not mean it is intended for the stage or exhibition booth. ## Can you get Memo for your event in Czech Republic and Europe? Not yet. Summary of the current situation: - **Event hire:** no, ROBOTio does not offer it, - **Sales and import:** no, ROBOTio is monitoring European availability, - **Europe:** Sunday is only discussing the USA at this stage, - **Intended use case:** household work, not events. Moreover, Memo is at a stage where its behaviour in household environments is being refined. For an event, you need something different: a robot that captivates guests from a distance, responds to interaction and fits into your programme. ROBOTio can notify interested parties if the situation changes, and ROBOTio can advise on cleaning robots for hotels and care facilities once they become available. An earlier overview is also available in the article [Sunday Memo: specifications, availability and roles for events](https://robotio.eu/blog/sunday-memo-domaci-robot-na-kolech-akce). ## What to deploy instead of Memo: Unitree G1 and H2 For events, ROBOTio has its own fleet of [Unitree](https://robotio.eu/brands/unitree) robots. The [Unitree G1](https://robotio.eu/unitree-g1-basic) at 135 cm tall (R.U.D.A. and R.I.C.O.) is complemented by the larger H2 at 180 cm. ROBOTio can deploy up to 4 of them at a single event. Each robot is accompanied by two operators throughout. What guests experience: - the robot speaks the language in which the guest addresses it (13 website languages), - you approve the character and dialogue in advance, - the robot maintains balance and movement independently, whilst the operator controls direction and gestures, - it can greet, host, dance and pose for photographs. A more detailed comparison of its full offer can be found in the [robot models](https://robotio.eu/models) overview. Typical applications are described on the pages [robot for corporate events](https://robotio.eu/robot-hire-by-event-type) and [robot for trade shows](https://robotio.eu/robot-hire-by-event-type). ### Proof in practice At the Brilon x Aquatherm trade show at PVA Expo Letňany, video from the robot's booth achieved over 2 million views, see [case study from Letňany](https://robotio.eu/case-study/pva-letnany-otevreni-haly). In Paris, three humanoids opened a company leadership meeting, see [TE Connectivity ALM Paris 2026](https://robotio.eu/case-study/te-connectivity-alm-paris-2026). ## How to proceed when planning your event 1. Describe your objective: welcoming guests, hosting, entertainment or content for social media. 2. Specify the location, date and number of guests. 3. Select a model: G1 for close interaction, H2 for greater impact on stage. 4. Approve the character and dialogue. 5. Receive a quote within 24 hours. If you are looking for a robot for cleaning in a hotel or care home, please contact ROBOTio as well. ROBOTio monitor the market and can advise you on what is currently feasible. ## Frequently asked questions ### Can you buy or hire Sunday Memo? No. Memo is not yet available for purchase. The beta programme for approximately 50 households in the USA is scheduled for the end of 2026, and ROBOTio does not offer it. ### Is Memo a humanoid? It has a human-like torso, head and two arms, but a wheeled column instead of legs and gripper hands instead of fingers. The manufacturer describes it as a domestic robot. ### Will Memo be available in Europe? Sunday has only announced availability in the USA so far. ROBOTio is monitoring European availability and will notify interested parties on its list. ### What does ROBOTio offer instead of Memo? Unitree G1 and H2 humanoids with two operators per robot, language-adapted character and deployment in Czech Republic, Slovakia, Europe and the Middle East. ## Do you want a robot for your event? Describe your event to ROBOTio and ROBOTio will respond within 24 hours with what ROBOTio can offer. Pricing is determined individually. Submit a non-binding enquiry through the [humanoid robot hire](https://robotio.eu/robot-rental) page. Want to see a humanoid at your own event? Find out how to rent a humanoid robot with its own operators. * Sunday Robotics * Sunday Memo * domestic robots * Unitree G1 * Unitree H2 * robot hire for events About ROBOTio. ROBOTio helps companies rent, demonstrate and integrate humanoid robots: Unitree G1 and H2 for events, trade fairs, conferences and marketing campaigns - always with its own operators.

Sunday Robotics debuts ACT-2 Preview, achieving 99.1% reliability for in-home AI robots. July 18, 2026 Reading Time: 3 minutes Key Takeaways: * Officially launched as a direct upgrade to November 2025's ACT-1, the ACT-2 Preview focuses strictly on reliability and long-tail environmental variations rather than controlled lab setups. * Operating on Sunday's home robot, Memo, ACT-2 achieved a 99.1% success rate during live testing, successfully completing 778 out of 785 autonomous laundry folds across unfamiliar households. * The model can absorb a brand-new behavior from just one fine-tuning demonstration and successfully propagate that skill across entirely new physical environments. User Score Sunday Robotics unveils the ACT-2 Preview, a groundbreaking robotics foundation model showcasing a 99.1% zero-shot success rate in folding laundry across completely unseen, real-world homes. Table of Contents The breakthrough mechanics behind ACT-2. The core engineering breakthrough of ACT-2 lies in its ability to combine broad, generalized environmental awareness with high-precision physical performance. Advanced pre-training framework. Sunday Robotics trained ACT-2 on a highly specialized dataset of sensorized human movement data. Following this initial phase, the model underwent automated, aggressive post-training loops on Sunday's flagship hardware, the Memo home robot. This data flywheel allows the system to achieve rapid performance gains, enabling the robot to master a completely new behavioral variation from a single fine-tuning prompt and replicate it seamlessly across foreign environments without requiring house-specific recalibration. Redefining industry success metas. To combat the industry trend of highly edited, single-take robotic video demonstrations, Sunday has introduced the "Solve" evaluation framework. This metric grades home robotics deployments across three distinct operational axes: Performance (the accuracy of the action), Scope (the diversity of environments the robot can handle), and Adaptation Cost (the amount of new data or human intervention required to launch the robot in a new location). Benchmarking performance data. To prove the operational maturity of ACT-2, Sunday selected laundry folding as its inaugural Solve target. The task is notoriously difficult for artificial intelligence due to the non-rigid, unpredictable nature of fabrics. Testing was conducted purely on unseen homes, cureal-world residential environments featuring unexpected lighting, varying floor layouts, and cluttered background spaces. The empirical results from the validation trials are detailed below: | Evaluation Category | Demonstrated Performance Metrics | | Overall Success Rate | 99.1% +/- 0.3% | | Total Autonomous Attempts | 785 trials | | Successful Task Completions | 778 folds | | Median Execution Time | 2 minutes and 13 seconds | | Average Fold Quality Score | 4.72 / 5.00 stars | The validation data shows that 98.3% of the completed folds achieved a 4 or 5-star rating under Sunday's strict internal quality rubric. Furthermore, the Memo robot successfully maintained this peak performance across a wide array of garment types, demonstrating advanced spatial reasoning by automatically recovering when clothing fell to the floor or when fabrics were heavily wrinkled and partially occluded. The future of consumer robotics. The launch of the ACT-2 Preview indicates that home robotics is moving closer to becoming a viable commercial consumer product. In traditional household settings, no two rooms, layouts, or items are identical, creating an infinite string of unpredictable edge cases known as the problem. By demonstrating that a robotic platform can enter an unmapped home and reliably handle complex, non-rigid object manipulation with almost zero failure, Sunday Robotics is proving that foundation models can successfully close the generalization gap. ACT-2 establishes a baseline where autonomous home assistants can finally function as dependable, practical tools for daily human life. Lindy McKenzie Lindy McKenzie brings over four years of experience analyzing the intersection of AI, blockchain, and global technology. She specializes in breaking down complex Crypto, AI frameworks into clear insights. Lindy helps readers navigate the legal complexities of emerging tech and data privacy with confidence.