Buddie: The Discreet AI Earbuds Assistant

Buddie’s Technology

Buddie’s Hardware Feature

Our Vision

Buddie is being developed by academic researchers in wearable electronics, AI, and human-centered computing. Steve Jobs revolutionized mobile phones by defining the touch screen as their primary interface. We believe that in the AI era, voice is the next transformative interface, and earbuds are the ideal form factor—enabling effortless access to AI services hands-free, anytime, and anywhere.

The team developing Buddie has the following goals: making AI services context-aware, and widely and seemlessly accessible in day-to-day life, giving innovators the freedom to experiment, and respecting user privacy. To this end, we are working with high-quality yet cost-efficient manufacturers and pricing Buddie at production cost. We’re also open sourcing Buddie’s mobile app to allow other developers to experiment and share new ideas and verify that Buddie respects privacy. Buddie is designed to protect private data, which doesn’t leave the phone unless explicitly requested by the user and users can easily delete their transcripts.

Open-Source Community

We’re primarily a team of researchers, although we have commercialized several technology products in the past (we were involved in the smartphone and wearables revolutions). We think efficient wearable AI interfaces will revolutionize computing and we want them to be open instead of walled gardens. Now is the time for invention and creativity so we’re focused on getting Buddie earbuds and its open-source app into many people’s hands as possible so they can experiment with it and try new ideas. We’re selling the Kickstarter product at production cost to make it broadly accessible. We’re also open sourcing the Buddie app, so developers can try and share new ideas for AI control and interaction. Designed for developers and users, Buddie offers unmatched flexibility, allowing customization through features like language model swapping. Through our open-source ecosystem and plugin marketplace, individual innovations become shared solutions that benefit the whole community. By contributing to Buddie, you’re helping to make open, privacy-focused, seemless, conversational AI assistance a reality.

Buddie Journey

We are a research team that has been deeply engaged in the areas of wearables and AI for 20 years. Over the past three years, we have been studying new ways to interact with AI. We’re now introducing an always-listening, privacy-first AI audio interface called Buddie. We started designing it in February 2024 and have been refining algorithms, software, and hardware to improve user experience.

Over the course of development, we made four major PCB revisions. By September 2024, we finally developed a stable system, which inspired us to design a brand-new headphone exterior specifically for it. By November 2024, we successfully completed the tooling process and received the first samples.

Media Coverage

Team

The Buddie team is a diverse group of researchers, engineers, and designers with decades of experience in wearable technology, AI/ML systems, and hardware design. We contributed to the smartphone and wearables revolutions, and have commercialized multiple products in these domains. Our academic contributions include over 300 research papers and eleven patents on AI, wearables, privacy, and special-purpose computing. Our goal is to make AI-powered devices open and accessible while keeping users in control of their private data, and to revolutionize human-AI interaction.

Robert: “Without context awareness, it’s like walking into a room where others are in the middle of a conversation. It’s hard to figure out what’s going on. If you were around the whole conversation, everything would make sense. Buddie uses patent-pending compression technologies to enable always-on operation so it can give AI assistants context. That’s the difference between asking about your life and asking what’s in an encyclopedia.

Buddie is designed for privacy. A system that transmits your voice to a remote server is safe only if its current and future owners are ethical. Buddie deletes audio data immediately after transcription, and they’re never transmitted to servers. Your transcripts and context information don’t leave your phone unless you specifically request it, for example to get help from a third-party AI, and you can delete them whenever you want.”

Yuhu: “We are working with the most reliable and cost-effective manufacturers in the world to produce Buddie. We have already determined Buddie’s electronics bill-of-materials, its molds are done, the printed circuit board is ready for manufacture, and we’re starting a small-scale run of hundreds of prototypes in two weeks.”

Jaric: “Buddie is an open and accessible platform. We want to make it easy for researchers and independent developers to try and share new ideas so we’re open sourcing the Budddie smartphone app. That also makes it easy for security researchers to verify that it protects privacy. We also want to make Buddie available to nearly everybody, so we’re setting the price so low that we expect to just break even on the Kickstarter campaign.”

Buddie team members have commercialized several products in the areas of wearable electronics and smartphones. We have a decade of experience with wearable hardware design and quality assurance. Some of us have been collaborating on research for 20 years, and some of us are still students. We have authored around 300 research papers on special-purpose computers and AI and have nine pending or issued patents on wearable AI.

Patents

Buddie’s team members have authored the following patents on wearable AI.

1. Voice Interaction Hardware-Software System Supporting Always-On Transcription, Real-Time Call Recording, and Automated Voice Command of AI Agents, 7 November 2024, 63/717,853, pending.
2. Context-aware Dialogue System, 23 October 2024, 18/924,655, pending.
3. A Privacy-Preserving, Low-Latency System for Continuous Audio Interaction with LLM-Based Assistants, 9 October 2024, 63/705,470, pending.
4. Understanding How People Read with Smart Eyewear, 30 May 2024, 63/469,706, pending.
5. Emotionship Analysis for Eyewear Devices, 26 January 2023, 18/101,856, pending.
6. MemX: An Attention-Aware Smart Eyewear System for Personalized Moment Auto-Capture, 26 September 2022, 17/952,810, pending.
7. Unveiling Causal Attention in Dogs’ Eyes with Smart Eyewear, 1 December 2022, PCT/CN2022/135926, pending.
8. Intelligent Scene Caching for Video Capture Data Reduction, 1 June 2021, 17/335,993, issued.
9. Digital Foveation for Low-Power Machine Vision, 25 September 2020, 17/032,499, issued.