> ## Documentation Index
> Fetch the complete documentation index at: https://docs.viorant.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Spaces

> An independent project where your prompts, skills, agents, models, and connectors work together as one runnable system.

A **Space** is an independent project. Your prompts, skills, agents, models,
connectors, and memory live together inside it as **real files** and work as one
system — you compose them into something that runs and ships.

Everything in a Space is on disk, in formats you can read, version, and move. The
Space is the container; the [Masters](/concepts/masters) are where you author each
part.

## What lives in a Space

* **Assets** — the things you author: prompts, skills, agents, and (soon) memory,
  each stored as its own file.
* **Pools** — the models and connectors this Space can draw on, imported from your
  [Model](/masters/model) and [Connector](/connectors/remote-mcp-oauth) setup.
* **A manifest** (`project.vpj`) — the Space's name, description, and default
  framework.

## Composition

The point of a Space is **composition**: separate, reusable parts assembled into a
working whole. The same prompt or model can back many agents, and each part stays
something you can reason about on its own. An [Agent](/concepts/agents) is the
assembly; the files are the source of truth.

## Portable by design

Each Space is versioned on its own and **compiles to a portable bundle you can run
locally now and deploy later** — so what you build stays yours and stays runnable.

<CardGroup cols={2}>
  <Card title="Create a Space" icon="folder-plus" href="/spaces/create-a-space">
    Start fresh or open an existing folder.
  </Card>

  <Card title="Add your first asset" icon="file-plus" href="/spaces/add-an-asset">
    Create a prompt, skill, or agent inside a Space.
  </Card>
</CardGroup>