Hugging Face Debuts DIY Humanoid Robots🤖

This week marks a significant leap across robotics, healthcare AI, and productivity tools. Hugging Face has introduced two open-source humanoid robots—HopeJR and Reachy Mini—designed to make robotics more accessible and customizable for developers, educators, and enthusiasts. Google followed with the debut of MedGemma, its most advanced open medical AI model collection yet, capable of interpreting both clinical text and medical imagery to support a new generation of digital health applications. Meanwhile, productivity takes a futuristic turn with the launch of Manus Slides, an AI-powered platform that generates complete, visually compelling presentations from a single prompt, streamlining content creation for professionals, educators, and teams everywhere. Together, these developments underscore the accelerating convergence of open innovation, domain-specific intelligence, and seamless AI integration into everyday workflows.

Hugging Face Debuts DIY Humanoid Robots

Hugging Face has unveiled two new open source humanoid robots, HopeJR and Reachy Mini, as part of its ongoing expansion into robotics. HopeJR is a full-size humanoid robot featuring 66 actuated degrees of freedom, allowing for complex movements like walking and arm motion, while Reachy Mini is a more affordable desktop unit designed for tasks such as moving its head, talking, listening, and testing AI applications. The company estimates HopeJR will cost around $3,000 and Reachy Mini between $250 and $300, with shipping expected to begin for initial units by the end of 2025 and a waitlist already open. These robots are designed to be accessible and modifiable, reflecting Hugging Face's commitment to open source robotics and reducing industry reliance on proprietary, black-box systems. The release follows Hugging Face's acquisition of Pollen Robotics and builds on recent initiatives like the LeRobot platform and the SO-101 robotic arm, furthering the company's mission to democratize robotics and AI development.

Google’s Advanced Open-Source AI for Medical Text & Images

MedGemma is Google’s most advanced open-source AI model collection for medical text and image comprehension, built on the Gemma 3 architecture and launched at Google I/O 2025. It comes in two main variants: MedGemma 4B, a 4-billion parameter multimodal model capable of processing both medical images (like X-rays, dermatology, ophthalmology, and pathology slides) and text, and MedGemma 27B, a 27-billion parameter text-only model optimized for deep medical text understanding and clinical reasoning. These models are designed as flexible foundations for developers to accelerate the creation of healthcare AI applications, including medical image classification, report generation, patient triage, clinical decision support, and summarization. MedGemma models are not clinical-grade out of the box; developers are expected to validate and adapt them—using prompt engineering, fine-tuning, or integrating with other tools—before production deployment. The models are accessible under open licenses for research and development, and can be run locally or on platforms like Google Cloud and Hugging Face, supporting innovation in digital health while maintaining privacy and control.

Manus Slides: Instant AI-Powered Presentations

Manus Slides is a newly released AI-powered tool that enables users to generate complete, visually appealing, and structured presentations instantly from a single prompt, leveraging advanced AI agent capabilities to optimize content layout and slide structure for diverse scenarios like business meetings, educational courses, and webinars. Users simply input their topic and goal, and Manus Slides automatically creates the entire deck, which can be easily edited by clicking on any section to personalize content, adjust layouts, or incorporate high-quality visuals and charts. The tool offers smart templates, adaptive learning, and customization options, making it accessible even for those without design skills, and supports exporting presentations in multiple formats such as PowerPoint, PDF, or images, with built-in sharing features for team collaboration and remote work. Manus Slides significantly boosts productivity by automating repetitive design tasks, providing real-time recommendations based on content, and allowing integration with documents or additional materials if needed. Its flexibility and efficiency make it ideal for professionals, educators, and content creators who need to produce high-quality, engaging presentations quickly, while also lowering the barriers to creation compared to traditional tools and setting a new standard for AI-driven productivity in presentation design.

Hand Picked Video

In this video, we’ll look at OpenManus, the ultimate open-source AI agent that lets you build and automate without restrictions, powered by GPT-4o and designed for seamless AI-driven workflows, from website creation to stock analysis, all completely free and accessible to everyone.

Top AI Products from this week

• Requirements Ai - Quickly convert your meeting transcripts into structured requirements in INCOSE, BDD, and User Story formats. Generate professional requirements in seconds.

• Knowledge - Discover high-intent accounts, qualify them with AI-powered insights, and contact the right decision-makers — all from one streamlined GTM workspace. No tabs, no guesswork. Just precise data, faster outreach, and smarter sales execution.

• VibeNecto - Generate professional marketing visuals for social media, ads, and more with our AI-powered platform. No design skills required. Our platform combines cutting-edge AI with intuitive design to help you create professional marketing content in seconds.

• Circuit Tracer - Anthropic's open-source Circuit Tracer helps researchers understand LLMs by visualizing internal computations as attribution graphs. Explore on Neuronpedia or use the library. Aims for AI transparency.

• ElevenLabs - The most realistic text to speech and voice cloning software. The most compelling, rich, and lifelike voices for creators and publishers seeking the ultimate tools for storytelling.

• Audino AI - Simplify video content creation with AI-powered audio generation. Our platform analyzes your videos to create perfectly synced sound effects and dynamic background music that adapts to every scene. Create content with ai audio that elevates your storytelling.

This week in AI

• Google AI Edge Gallery - Run Hugging Face AI models fully offline on Android. Experimental alpha, Apache 2.0 license. Latest release: GitHub, May 21, 2025.

• QwenLong-L1 - Long-context LLM with RL, supports up to 131K tokens, outperforms top models. Apache-2.0 license. Latest update: GitHub, May 28, 2025.

• AI influencers - Virtual creators like Imma and Aitana attract followers and brand deals, reshaping social media with lifelike avatars, interactive chatbots, and AI-driven marketing.

• One-App AI Phone Future - Carl Pei envisions future smartphones with just one app—the OS—powered by AI to automate tasks and personalize user experience.

• Windows Unveils New AI Tools - Microsoft announced Windows AI Foundry, new APIs, and native support for AI agents, making Windows 11 a top platform for AI development and agentic apps.