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

# Core concepts

> The eight terms you need to read the rest of the docs: project, application, repository, sandbox, thread, agent run, workflow and environment.

Revolte uses a small set of terms the same way on every page. Each entry below gives a short definition and a link to the page that goes deeper. The full list is in the [glossary](https://docs.revolte.ai/concepts/glossary.md).

## Project

A project groups related applications. Your plan sets how many projects and applications you can have. See [Pricing and usage plans](https://docs.revolte.ai/pricing.md).

## Application

The thing you build, deploy and operate in Revolte. An application is connected to a repository source and has its own environments and configuration. You can move an application to a new repository without affecting active deployments. See [Migrate existing repositories](https://docs.revolte.ai/integrations/repository/migration.md).

## Repository

The code Revolte reads, branches and opens pull requests against. Revolte works with GitHub repositories today. See [Connect GitHub](https://docs.revolte.ai/integrations/repository/github.md).

## Sandbox

An isolated cloud environment where the agent reads, writes and runs your code. Every task runs in a sandbox, so the agent never works directly on your machine or on production. See [Sandboxes](https://docs.revolte.ai/get-started/sandboxes.md).

## Thread

One conversation with Revolte about a task in Interactive Chat. You can return to a thread later to continue the work. See [Continue existing work](https://docs.revolte.ai/workflows/continue-previous-work.md).

## Agent run

One execution of the agent against a task, for example an autonomous workflow run that takes a Jira ticket to a pull request. A run keeps its plan, code changes and status so you can review what happened. See [Review the execution plan](https://docs.revolte.ai/workflows/review-execution-plan.md).

## Workflow

A workflow coordinates agents, context, a sandbox, human checkpoints and external systems around one outcome. It can start from a prompt, a Jira ticket, a bug report or a repository. You can run one interactively or hand it off and come back to review. See [Workflow overview](https://docs.revolte.ai/workflows/overview.md).

## Environment

Where an application runs: Production, Staging, QA or Preview. Each environment has its own settings, variables and limits. Preview environments let you validate a change before you merge it. See [Environments and promotion](https://docs.revolte.ai/deployments/overview.md).

## How the terms fit together

You connect a **repository** to an **application**. You start a **thread** or a **workflow**. Each task runs in a **sandbox** and produces a pull request. Changes move through **environments** until they reach Production. Where a person approves is described in the [human-in-the-loop model](https://docs.revolte.ai/concepts/human-in-the-loop.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.