> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.cloudraker.com/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.cloudraker.com/_mcp/server. # What CloudRaker is CloudRaker is a platform where you and a built-in AI assistant do repetitive, document-heavy office work together. Examples: intake, invoice processing, contract review, vendor onboarding, and claims. Your work lives in a **space**. An admin defines a reusable process once as a **playbook**. The AI assistant runs the playbook and drafts the outputs. It reads documents, fills forms, and drafts records. **You always review and approve before anything becomes real data.** ## Who it's for CloudRaker serves business and operations teams that run document-heavy, multi-step processes. You do not need to be technical. If your job is to collect files, pull data out of them, fill forms, and produce documents, CloudRaker is for you. ## The one rule that defines CloudRaker **The assistant drafts; you decide what to approve.** The AI assistant and automated steps never commit data on their own. They produce drafts: proposed records, filled forms, and generated documents. A person reviews, edits, approves, or rejects before anything becomes committed data. Every extracted value, every filled form, and every record starts as a draft. ## How the pieces fit together Your **organization** is your whole workspace. Inside it, work happens in **spaces**. A space is a workspace that holds files, records, and people for one body of work. Every space is created from a **space type**, a reusable template that an admin sets up. Admins also define reusable **playbooks** (an ordered process the agent can run). Admins install **actions** (single-purpose capabilities such as extract, fill a form, or sign). When you run a playbook inside a space, you get a **run**. The agent works through the run's tasks and produces drafts that you approve. How CloudRaker fits together: space types, playbooks, and actions are set up once in the organization; spaces are created from them; running a playbook in a space produces drafts that a person approves into records, documents, and files Keep these relationships straight: * A **space** is where work happens. A **space type** is the template a space is created from. Admins manage space types. You work in spaces. * A **playbook** is a blueprint defined once at the org level. A **run** is one execution of that playbook inside a space. You can run one playbook many times. * The **agent** does the drafting inside a run. A person approves before any AI-produced value is committed. For a one-line definition of every term (space, space type, playbook, run, record, action, grounding), see [Key concepts](/workspace/get-started/concepts). ## Signing in Sign in to CloudRaker at [app.cloudraker.com](https://app.cloudraker.com) with your **work account**. Your organization's identity provider handles sign-in. There is no local CloudRaker password to set or reset. If you visit a page while signed out, the app sends you to sign in. It then returns you to the page you requested. The hosted sign-in screen with email, Google, and Microsoft options ## One organization, one tenant Everything you see belongs to a single **organization**: spaces, files, records, and roles. If you are a member of more than one, switch between them from the switcher at the top of the sidebar. Switching re-scopes the entire app to the other organization and reloads it. Nothing carries across. ## Where to go next #### [Quickstart](/get-started/quickstart) Sign in, upload files, run a playbook, and approve the agent's work. #### [Key concepts](/get-started/concepts) Plain-language definitions of every term you will meet in CloudRaker. #### [Spaces](/spaces/overview) See what a space is and how its tabs are organized. #### [Playbooks](/playbooks/overview) Understand the blueprints that drive your repeatable processes.