> 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 RakerOne is RakerOne is a back-office automation platform where you and a built-in AI assistant, **Florent**, work together to get repetitive, document-heavy office work done — intake, invoice processing, contract review, vendor onboarding, claims, and the like. You define a process once as a reusable **playbook**, then run it inside a **project** that holds your files, records, and people. Florent does the heavy lifting — reading documents, pre-filling forms, drafting records and documents — but **you always review and approve before anything becomes real data.** ## Who it's for RakerOne is built for business and operations teams who run document-heavy, multi-step processes and want AI to remove the grunt work without giving up control. You do not need to be technical. If your job involves collecting files, pulling data out of them, filling forms, and producing documents — over and over — RakerOne is for you. A typical moment: a team drops 400 PDFs into a project and types *"Extract these into my Invoice table."* Florent turns that sentence into a concrete run, you review the shape and start it, the documents get read by AI, and every extracted value lands as a draft for you to check. You approve, edit, or reject each row — and only then does it become real data. ## The one rule that defines RakerOne **AI drafts, humans approve.** Florent and automated steps never commit data on their own. They produce drafts — proposed records, pre-filled forms, generated documents — and a person reviews, edits, approves, or rejects before anything becomes committed data. This is the promise behind everything in the product. Florent will **never**: * activate or publish a draft, * approve or reject a task, * submit or complete a task for you, * start or cancel a run, * delete anything, or * set who can access something. You always perform the final, consequential action yourself. Florent shapes, reads, and proposes; you decide. ## How the pieces fit together Hold this one picture in your head and the rest of the app makes sense. Your **organization** is your whole workspace. Inside it you keep two things: **Building Blocks** (reusable definitions you set up once — project templates, playbooks, actions, record types, document templates) and **projects** (where the actual work happens, one project per piece of work). When you run a playbook inside a project you get a **run**, which is made of **tasks** — the individual steps a person or Florent works through. As tasks finish, they produce **records** (rows of data) and files. A human approves before any AI-produced data is committed. A few relationships worth fixing in your mind: * A **playbook** is a blueprint. A **run** is one execution of that blueprint inside a project. You can run one playbook many times. * A **record type** defines the shape (the fields). A **record** is one row of that shape. Approved drafts become records. * A **task** is one step in a run. People do **Form** and **File upload** tasks; Florent and automation handle **Document**, **Action**, and **AI task** steps. * **Project templates are snapshot-copied** — creating a project from a template copies its contents, and later edits to the template do not change projects you already created. This page gives you the mental model. For a one-line definition of every term — playbook, run, record, brief, grounding, and the rest — see [Concepts](/get-started/concepts). ## A tour of the app Every screen sits inside the same frame: a **left sidebar** for navigation, a top **breadcrumb trail** showing where you are, and the main content area on the right. ### The sidebar, top to bottom The top of the sidebar shows your current organization's name and logo. If you belong to more than one organization, click it to switch — switching reloads the whole app so nothing carries over between organizations. [Learn more](/admin/organization-settings). Three items sit at the top: * **Home** opens [My work](/work/my-work), your personal list of tasks waiting on you across all projects. * **Inbox** opens your notifications — mentions, task assignments, and approval requests. It pops over the current screen as an overlay rather than opening a page, and a badge shows your unread count. [Learn more](/work/collaboration). * **Chat** is where you talk to Florent. [Learn more](/work/florent). Under the **Projects** label, your open projects are listed (up to 10, then a **Show more...** link). A **+** button next to the label starts a new project — you only see it if you can create projects. You can drag projects to reorder them. The **Building Blocks** group holds the reusable definitions you set up once and reuse: **Project templates**, **Playbooks**, **Actions**, **Record types**, and **Document templates**. You only see this group, and only the items within it, if you have the matching permissions. The **Admin** group — **Settings**, **Users**, **Roles**, and **API keys** — appears only if you're an organization admin. The bottom of the sidebar shows your name. Open it to switch the interface [language](/admin/language) or to log out. The Building Blocks links are grouped under **Building Blocks**, not Admin. If you're looking for Playbooks or Record types, go to **Building Blocks** — the **Admin** group is only for managing people and the organization. ### The breadcrumb The top bar shows a breadcrumb trail of where you are — for example, **Projects · Smith family · Files**. It always shows human-readable names (project names, playbook names, run numbers), never internal IDs. The breadcrumb is hidden on the Home page, which is its own starting point. Your sidebar may look different from a teammate's. Items appear and disappear based on what your role lets you do, so two people in the same organization can see different menus. See [Roles and permissions](/admin/roles-and-permissions). ## Where to go next The fastest path to value — create a project, start a playbook, and see records appear. Plain-language definitions of every term you'll meet in RakerOne. Find and claim the tasks waiting on you across every project. See what a project is and how its tabs are organized. Understand the blueprints that drive your repeatable processes.