The operating system for AI-native software.
Four products, one continuous fabric. From first prototype to live, monetised application without ever leaving the ecosystem.
BoltzHub
An innovative, "vibe-coded" marketplace where creators can sell AI-powered apps and consumers can easily subscribe to them. Every app features advanced capabilities like agent chat, workflow automation, and productivity tools, all backed by a scalable backend that processes billions of records and files. Powered by Boltzbit live learning technology, these apps dynamically adapt to user behavior to continuously improve accuracy and efficiency.
Explore
BoltzVibe
A specialized vibe coding agent designed to develop apps with seamless integration into the Boltzhub ecosystem. It features the Boltzcode CLI, an intuitive command-line tool that allows users to write and build code entirely through chat. For a seamless developer experience, it also includes a VS Code plugin, bringing the power of the Boltzcode CLI directly into your favorite IDE.
BoltzLearn
An advanced LLM fine-tuning and live learning platform that empowers users to build custom AI models using their own data. Once trained, users have the flexibility to download their models for self-deployment or host them directly on Boltzhub or BoltzCloud. It provides the ultimate control for businesses looking to create tailored, continuously learning AI solutions.
BoltzCloud
A virtual cloud wrapper that simplifies multi-cloud deployment across GCP, AWS, and Azure without complex networking or unpredictable costs. It serves as the native, pre-configured hosting backbone for all Boltzhub apps to ensure seamless scalability. With BoltzCloud, startups and enterprises can easily deploy their AI apps anywhere with fine-grained resource control and superior price-to-performance than direct cloud hosting.
What AGI 2.0 stands for
Standard 1:
Cost-efficient models learning during inference
Today, the costs of developing AGI systems are prohibitively high. To overcome this hurdle, models must be able to learn and adapt while they are being used (during inference), and do so without dramatically increasing costs. This shift from massive one-off training runs to continuous, lightweight learning would break the current economic bottleneck.
Standard 2:
Full ownership and control of models and data
The lack of clear ownership and intellectual property rights is a major barrier to progress. If the goal is truly to empower people, they must be able to own their machine learning model outright, not just rent access from someone else.
They must also own and control the data they use to train or fine-tune that model, including where it is stored, how it is used, and who can access it. And they must retain the intellectual property created through their models and data, ensuring that private training data doesn’t simply become free fuel for someone else’s platform.
Standard 3:
Dynamic interface powered by natural language
Today’s static, pre-designed software interfaces are another barrier to AI adoption. Users must learn the tool before they can benefit from it.
By contrast, a dynamic, natural-language-driven interface would let users simply describe what they want in plain language and have the system build or adapt the interface on the fly. In short, AI should adapt to the user, not the other way around.
The foundations of an open intelligence era
AGI 2.0 is designed to be owned, shaped, and governed by the many, not the few.
Sustainable intelligence creation
AGI shouldn’t cost billion-dollar to train and serve.
AGI 2.0 makes intelligence affordable by enabling models to learn during inference, not through endless retraining cycles. This keeps innovation open, competitive and economically accessible.
Ownership and
sovereignty
True empowerment requires true ownership.
AGI 2.0 protects the right to build and fully own your intelligence without feeding it into someone else’s platform.
Adaptive and interoperable intelligence
Intelligence should evolve with you.
AGI 2.0 replaces static plug-ins and siloed tools with systems that adapt in real time, reason across contexts, and connect to each other seamlessly.
Impact
Why it matters
AGI 1.0 built on scale and static training has reached its limits. Its economics don’t work. Its power is too concentrated. And it leaves most of the world in the role of users, not owners.
AGI 2.0 is a course correction.
It turns intelligence into infrastructure and puts its creation back in their owner's hands. If we get this right, AGI will become the great equaliser of the Intelligence Age, not its gatekeeper.
"We must ensure that intelligence itself remains a public good, not a private empire."
Dr. Yichuan Zhang
CEO & Chief Research Officer
Our values
Our values reflect the future we’re working towards for our team, our technology and the world we serve.
We build with purpose
Every product we create is designed to solve real problems, deliver meaningful impact, and move the world forward responsibly.
We drive impact
We measure success by the value our technology creates, not the noise it generates.
We simplify
We design technology that feels natural and intuitive.
Experience
Our teams
We are researchers, engineers, and operational leaders, with experience spanning a variety of disciplines, all working together to build reliable GenAI systems.
Research
Our research team, composed of leading experts in machine learning, computer vision, and natural language processing, is dedicated to advancing the forefront of GenAI technology.
Business
Our business team drives strategic growth and fosters partnerships, ensuring we create value and deliver innovative solutions to our clients.
Product
Our product team is focused on designing intuitive, high-impact solutions that meet the evolving needs of our users and shape the future of AI technology.
Engineering
Our engineering team builds robust, scalable systems that power the next generation of AI, ensuring reliability, performance, and cutting-edge capabilities.
