> ## Documentation Index > Fetch the complete documentation index at: https://docs.rippit.com/llms.txt > Use this file to discover all available pages before exploring further. # Google Drive > Let the Rippit agent search, read, create, update, and share Google Drive files. Google Drive connects through Rippit's Google app. You do not need a Google Cloud project, OAuth client, Developer Preview enrollment, or Service Usage Consumer role. ## Connect 1. In **Connections → MCP Servers**, choose **Google Drive**. 2. Choose who can use the connection. Organization-wide access is selected by default; if you keep it selected, acknowledge that all members and Agent Apps will act through the Google identity that signs in. 3. Review the Workspace data-use disclosure, acknowledge it, then continue to Google and approve the requested access. 4. After Google sign-in, use **Add Drive items** to choose individual files or folders. An individual file is available to the agent. A selected folder is available as a destination for new items, but it does not grant read access to the files already inside it; every existing file the agent should read has to be selected individually in Picker. The admin who created the connection is the one who can open Picker, change selected items, and disconnect it. Other admins see the connection as view-only: they can review which items are selected but cannot add or remove them. Workspace members can use the selected items through enabled tools but do not receive the Google token or Picker credentials. Google grants the Rippit OAuth app the non-sensitive `drive.file` scope. Rippit and Google both limit access to items selected through Picker and items Rippit creates. Rippit also applies its own positive allowlist before every tool call. Removing an item from the connection removes it from that Rippit allowlist. Files and folders Rippit creates are added to the allowlist automatically. This lets an agent continue working with a created item later without granting access to unrelated Drive content. ## What the agent gets * Search and recent-file listing. * Metadata, permissions, and comments. * Structured text from Google Docs, Sheets, and Slides. * Text from UTF-8 text, CSV, JSON, XML, HTML, and PDF files. * Raw downloads, including files whose formatted content Rippit cannot render. * File creation, copying, and sharing when an admin enables the corresponding write tools. * Google Sheets updates for values and formulas by A1 range, including appended rows. * Atomic Google Docs updates for inserting, appending, replacing, formatting, and deleting content. Word, Excel, PowerPoint, OpenDocument, and image files remain downloadable but are not converted into formatted text. The agent reports that limitation and can use `download_file_content` when raw bytes are appropriate. Rippit will not read trashed files, files Google marks as non-downloadable, or client-side encrypted files it cannot parse. It does not bypass Google's malware or abuse warnings. Individual downloads and parser inputs are limited to 10 MiB; rendered text is limited to 200,000 characters and reports when it was truncated. ## Tools The connection exposes a fixed twelve-tool contract. The six read tools are enabled initially; write tools remain off until an admin enables them in **Manage tools**. | Tool | Access | What it does | | --------------------------- | ------ | --------------------------------------------------- | | `search_files` | Read | Search with Google Drive API query syntax | | `list_recent_files` | Read | List recent files | | `get_file_metadata` | Read | Read metadata for a known file | | `read_file_content` | Read | Render supported content and optional comments | | `download_file_content` | Read | Download or export content as base64 | | `get_file_permissions` | Read | List a file's permissions | | `update_spreadsheet_values` | Write | Overwrite cells by A1 range with values or formulas | | `append_spreadsheet_values` | Write | Insert rows after an existing table | | `update_doc` | Write | Apply atomic Google Docs batch updates | | `create_file` | Write | Create or upload a file | | `copy_file` | Write | Copy a file | | `create_file_permission` | Write | Share a file with a user, group, domain, or anyone | ## Data handling Rippit uses your Google Workspace content only to answer the chat, run the Skill, or run the Action you asked for, and may send that content to the model provider behind that feature. Rippit never uses Workspace content to train or improve AI or ML models. Treat everything in Drive as untrusted text. If a file or a comment contains instructions, the agent will not follow them. They cannot trigger tool calls or override what you or Rippit told the agent to do. ### What Rippit stores Rippit does not store raw file content pulled from Drive. Content is kept only for as long as the request that reads it, and Rippit does not cache Drive files. What is retained after a request finishes: * The connection settings, including the IDs of the files and folders selected in Picker and the IDs of items Rippit created. * File metadata such as IDs, titles, size, and timestamps. * The text input by the user and the response from Agentic Chat, which can include snippets from referenced documents. That last one is the only place Drive content itself can persist. Anything the agent quotes or summarizes from a file is stored with the chat, like any other chat message. ## Troubleshooting **Picker did not open or failed to add items:** use **Add Drive items** on the connection card. If that also fails, contact Rippit support. **A selected file is not visible:** reconnect if Google access was revoked, then select the file again. The connected Google account must still be able to open it.