The co-founders behind viral apps Reface and Prisma have unveiled Mirai, a new venture focused on optimizing AI model inference directly on mobile devices. This initiative aims to enhance performance, efficiency, and data privacy by moving AI processing from the cloud to local hardware. Leveraging their expertise in deep learning for image and video manipulation, the team targets faster, more private AI applications across multiple sectors.
In a strategic move that aligns with the accelerating shift toward on-device AI, the masterminds behind two of the most visually captivating mobile applications—Reface for face-swapping and Prisma for artistic photo filters—have joined forces to announce their ambitious new project: Mirai. This venture directly tackles one of the most significant technical challenges in contemporary artificial intelligence: optimizing model inference to run efficiently directly on smartphones and personal computers, eliminating the need for constant cloud connectivity. By merging their pioneering expertise in neural network-based image and video processing, the collaboration represents a direct response to the growing market demand for AI applications that are faster, more private, and less power-hungry, fundamentally changing how users interact with intelligent technology.
The Mirai project is a technical initiative focused on developing advanced software and libraries to streamline the execution of AI models, particularly complex deep learning models, directly on device processors. The core philosophy is to empower developers to build applications that deliver instant responses and handle user data with complete privacy, as all processing occurs locally without transmitting sensitive information to external servers. This approach not only enhances security but also reduces latency, creating a seamless user experience for real-time applications.
The founding team is building upon the immense lessons and technical prowess gained from their previous successes. Prisma, which amazed users by applying the styles of master painters to photos in near real-time, and Reface, which popularized hyper-realistic face-swapping in videos and GIFs, both rely fundamentally on sophisticated AI architectures. Their hands-on experience in optimizing these demanding models for a wide range of consumer devices with limited computational resources forms the foundational expertise that the Mirai project will leverage to push the boundaries of local AI execution.
Mirai's solutions are expected to target a broad spectrum of industries and use cases, including:
The launch of Mirai arrives at a pivotal moment as the tech industry undergoes a fundamental paradigm shift from centralized cloud computing toward distributed edge computing. While the cloud offers immense scalable power, it inherently suffers from issues of latency, data privacy concerns, and the energy cost of constant data transmission. On-device AI solutions like those promised by Mirai address all three of these shortcomings simultaneously, offering a more sustainable and user-centric model for the next generation of intelligent apps.
From a competitive standpoint, the Reface and Prisma founders are entering a market increasingly crowded with tech giants like Google (with TensorFlow Lite), Apple (with Core ML), and Qualcomm (with its AI accelerator platforms). However, their unique pedigree in creating visually stunning consumer applications that went viral, combined with a proven ability to simplify complex AI technology for mainstream developers, could carve out a distinct niche. Their prior success in turning advanced AI into social media trends demonstrates a deep, practical understanding of end-user desires that purely infrastructure-focused competitors may lack.
Model inference is the phase where a pre-trained artificial intelligence model is used to make predictions or decisions based on new, unseen data. For example, when you upload a photo to the Prisma app, the process that transforms it into a Van Gogh-style painting is an inference task. Mirai's goal is to make this process significantly faster and more energy-efficient when running directly on your phone's processor instead of a remote server, enabling instant creativity and interaction.
The primary advantages are threefold: Speed (dramatically reduced latency for real-time interaction), Privacy (sensitive user data never leaves the personal device), and Reliability (functionality continues uninterrupted without an internet connection). Additionally, it reduces bandwidth costs for users and developers and can lead to better overall battery life by minimizing constant data uploads and downloads.
Mirai focuses on overcoming the challenges of running large, complex neural networks on hardware with limited memory and processing power. This involves techniques like model optimization, quantization (reducing the precision of calculations to speed them up), and efficient scheduling of tasks on a device's CPU, GPU, or dedicated AI accelerator. Their expertise in making photo and video AI work smoothly on phones is directly applicable to this broader problem.
The primary audience is mobile and software developers looking to integrate advanced AI features—like real-time image enhancement, object recognition, or natural language processing—into their applications without compromising on speed or user privacy. By providing optimized tools and libraries, Mirai aims to democratize access to high-performance on-device AI, making it easier for developers of all skill levels to build the next generation of intelligent apps.
While frameworks like TensorFlow Lite provide the foundational tools for on-device inference, Mirai appears to be building a higher-level platform or suite of pre-optimized solutions, potentially focusing on specific verticals like media processing. Their key differentiator is the applied experience of taking cutting-edge AI research and productizing it into wildly successful consumer apps. They are likely to offer not just tools, but also best practices and architectures honed through real-world, mass-market deployment.
The announcement of Mirai marks a significant evolution in the AI landscape, signaling a move from cloud-dependent intelligence toward more personal, immediate, and private computing. By leveraging their unparalleled experience in building AI-powered apps that captured the global imagination, the founders of Reface and Prisma are uniquely positioned to accelerate the adoption of on-device inference. If successful, Mirai could empower a new wave of applications that are not only smarter but also more respectful of user data and responsive to our needs in real-time, ultimately making advanced artificial intelligence a more integrated and seamless part of our daily digital lives.
Source: TechCrunch AI | Analysis & Editorial: AI Tools Oasis
Bringing you the latest news and analysis in the world of Artificial Intelligence with accuracy and credibility. Follow us for all updates.
OpenAI is advancing its ambitious super app project, aiming to integrate advanced AI capabilities into a single, multifunctional platform. This development is part of the company's strategy to expand services and deliver a unified user experience. Discover the full details and expected impact of this move.
Notion has restored access to its Anthropic AI integration after a 4-hour outage disrupted users relying on Claude-powered features. The incident highlights the growing dependency on AI productivity tools and raises questions about infrastructure stability. All user data remained secure during the disruption.
A new report from TechCrunch AI warns of a potential 'Tokenpocalypse'—a massive collapse of digital tokens due to oversupply. With over 80% of new tokens losing 90% of their value, the market faces a crisis reminiscent of the dot-com bubble. This analysis explores the risks, impacts, and how investors can protect themselves.
