> For the complete documentation index, see [llms.txt](https://docs.ua1.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ua1.ai/overview/what-is-ua1.md).

# What is UA1?

What is UA1?

UA1 is a programmable infrastructure for deploying persistent, intelligent, and economically active agents. These agents are not chatbots or APIs — they are autonomous labor units that reason, evolve, and operate across real-world platforms.

Each agent is fully encapsulated with identity, memory, skills, and behavioral logic. It can think, act, earn, and improve over time — without requiring human micromanagement.

<figure><img src="/files/VKtLH0pOkMg5Uv6kUQVs" alt="" width="563"><figcaption></figcaption></figure>

#### Why UA1?

Because for the first time in human history, labor can be manufactured — trained, owned, deployed, and monetized at scale. UA1 transforms AI from a service into **synthetic capital** — programmable, composable, and sovereign.

#### Multi-Compute by Design

UA1 introduces **MCP (Multi-Compute Provider)** architecture to ensure every agent can run anywhere:

* Secure enclaves for sensitive missions
* Edge or on-prem for data locality
* Distributed cloud for high throughput

Each agent selects or is assigned a compute runtime based on:

* Mission context
* Security level
* Owner preference

<figure><img src="/files/gcu6bJiBqIBdg0s7Ih6Q" alt=""><figcaption></figcaption></figure>

UA1 is not cloud-AI. It is **compute-agnostic synthetic labor** — modular, auditable, and sovereign by design.