Startup offers free home cleaning—if it can record it all for robot training
A new startup, MicroAGI, is offering free home cleaning services in exchange for recording the cleaning process to train AI-driven robots. The company claims to anonymize sensitive data, but questions remain regarding data removal and potential identification of homes. In parallel, the app recruits individuals to record daily tasks, paying them to generate valuable robot training data.
MicroAGI, a German startup, has launched the Shift app, offering free home cleaning services to New York City residents. The unconventional offer requires "professional cleaners" to wear cameras and record the entire cleaning process. This data is intended to train AI-driven robots, accelerating the development of embodied AI. The company promotes this service as a "no-catch" deal, despite some hidden conditions.
The Shift app's website directs users to book appointments and provide personal information. While the company claims to automatically anonymize sensitive details, blurring faces and identifiers, it does not specify if users can request the removal of their home cleaning videos from training datasets. This raises privacy concerns about potential identification of homes within the datasets.
Booking a free cleaning through the Shift app requires payment information. Clients may be charged for cancellations made with less than 24 hours' notice or if they are unavailable for their appointments. The terms of service also absolve the platform of responsibility for any property damage, theft, or personal injury that may occur during the cleaning.
The initial free cleaning offer serves as a promotional tool for the app's main purpose: recruiting individuals to record everyday tasks for a fee. The Shift app seeks "operators" to wear recording headstraps and capture videos of household or professional activities, paying them up to $20 per hour plus bonuses. The company claims to have paid over $5 million to more than 10,000 operators in the first quarter of the 2026 fiscal year.
MicroAGI is actively recruiting in various communities, including university students, teachers, and restaurant workers in NYC, through targeted blog posts and Craigslist advertisements. The company also plans to expand its services to other major cities like London, Munich, and Zurich. This strategy highlights a growing trend among startups to collect real-world data from ordinary people to advance robot training.
Other companies like Encord and Micro1 are also engaging thousands of contract workers globally to gather data for robot training, emphasizing the increasing demand for diverse and realistic datasets in the field of AI and robotics.
Related articles
Build real agentic apps using CUGA: two dozen working examples on a lightweight harness
CUGA, IBM's open-source Agent Harness, simplifies building agentic applications by handling infrastructure, allowing developers to focus on tools and prompts. It offers pre-assembled components for planning, execution, and state management, significantly reducing development time. CUGA has topped agent benchmarks like AppWorld and WebArena.
OpenAI launches new initiative to help find and patch open source bugs
OpenAI has launched "Patch the Planet," a new initiative in partnership with cybersecurity firm Trail of Bits, to enhance the security of open-source projects. This program aims to assist maintainers in identifying and patching bugs, utilizing OpenAI's AI-powered security tools while reducing the burden on project teams.
PP-OCRv6 on Hugging Face: 50-Language OCR from 1.5M to 34.5M Parameters
Baidu has released PP-OCRv6, an advanced optical character recognition (OCR) model supporting 50 languages. Available on Hugging Face, this version significantly improves accuracy and efficiency across various parameter sizes, from 1.5 million to 34.5 million, marking a substantial leap in multilingual OCR technology.