--- title: "Outline Integration" description: "Connect AI agents to Outline, your team knowledge base. Agents create, update, and full-text search wiki documents, append meeting notes and runbooks automatically, browse collections, and look up workspace users." url: https://flowrunner.ai/integrations/outline date_modified: 2026-08-01T02:40:32-07:00 --- # Outline [Productivity](https://flowrunner.ai/integrations/category/project-management-productivity) Connect AI agents to Outline, your team knowledge base. Agents create, update, and full-text search wiki documents, append meeting notes and runbooks automatically, browse collections, and look up workspace users. [Verified](https://flowrunner.ai/integrations/verified "What does verified mean?") · 10 actions · API key · available [Outline website](https://getoutline.com/) · [Platform Documentation](https://www.getoutline.com/developers) · Capability data verified 2026-07-28 1. An incident closes in the on-call system with a resolution summary attached 2. Search Documents checks whether a runbook already covers this failure mode 3. If one exists, Update Document with Append adds the incident's date, symptoms, and fix to its history section; if not, Create Document drafts a new runbook in the Engineering collection as an unpublished draft 4. The agent verifies the draft's structure: symptoms, diagnosis steps, fix, and links to the incident record 5. The on-call engineer reviews the draft for technical accuracy and publishes or corrects it 6. The published document link posts to the incident channel, closing the loop 7. The incident record gets the runbook link written back for the next responder ## What This Integration Enables A knowledge base is only as good as its worst habit, and every team's worst habit is the same: writing things down is nobody's job at the moment knowledge is created. FlowRunner makes it an agent's job. Agents write into Outline from the systems where knowledge actually originates, incidents, meetings, releases, support escalations, using Markdown documents created in the right collection or appended to the right page. They read back out with full-text search that returns ranked matches with highlighted snippets, which turns the wiki from a place people forget to look into a source agents consult mid-workflow. - Append meeting notes, changelog entries, and incident histories to living documents automatically - Draft new runbooks and docs as unpublished drafts, with a human publishing after review - Search the wiki inside workflows and deliver the matching document into the channel where it is needed - Inventory documents, collections, and users on a schedule for audit and sync - Gate full-body replacement and permanent deletion behind a person ### Without FlowRunner **Knowledge retires with the incident**: The fix lives in a chat thread that scrolls away, and the next occurrence starts from zero **The wiki decays silently**: Runbooks describe the architecture of two migrations ago, because updating docs loses to shipping **Search happens in people's heads**: Finding the right doc means asking the person who wrote it, if they still work here ### With FlowRunner **Documents accumulate from events**: Meeting notes, changelogs, and incident histories append themselves as the events happen **Drafts arrive for review, not from scratch**: Agents produce structured drafts and humans spend their attention on accuracy, the part only they can supply **Search is a workflow step**: Full-text search runs inside automations, so the relevant runbook surfaces in the incident channel before anyone asks ## Use Case Scenarios ### The runbook that gets smarter after every incident An incident closes. Search Documents finds the matching runbook, and Update Document with Append adds a dated entry: what broke, what the responder saw, what fixed it. Six incidents later, the runbook's history section reads like the diagnostic flowchart nobody had time to draw. When a new alert fires, the workflow searches Outline and posts the runbook link with its snippet into the [Slack](https://flowrunner.ai/integrations/slack) incident channel before the responder finishes acknowledging. ### Meeting notes that file themselves After each leadership meeting, the workflow takes the transcript summary, formats decisions and owners as Markdown, and appends to the quarter's running decisions document in the Leadership collection. Anyone asking "when did we decide that" gets an answer from Search Documents instead of an archaeology dig through calendars. The document's edit history is the decision log. ### The migration that moved a wiki without a lost page A team consolidates from [Confluence](https://flowrunner.ai/integrations/confluence) into Outline. An agent reads each source page, converts it, and republishes via Create Document into the mapped collection as unpublished drafts, batch by batch. List Documents inventories progress into [Google Sheets](https://flowrunner.ai/integrations/google-sheets), and a reviewer publishes each batch after spot-checking. The migration takes evenings, not a quarter, and the audit sheet shows exactly which page became which document. ## Human-in-Loop Highlight Update Document replaces the entire body by default; Append is the opt-in. That one default is this connector's sharpest edge, because the failure is silent: an agent that intends to add a paragraph but omits the Append flag publishes a one-paragraph document where a two-hundred-line runbook used to be, and everything looks fine until 3 a.m. when a responder opens it. Delete Document has a matching trap at the end of the lifecycle: the default delete is a restorable trash move, but with Permanent enabled the document skips the trash and is gone irreversibly. FlowRunner workflows are built against both edges. Automated writes to existing documents run in Append mode; a full-body replacement must be explicitly requested and shows the human the before and after. Permanent deletion never runs on an agent's initiative: a cleanup workflow may trash, but only a person who has seen the document's title, collection, and content confirms the version of delete that forecloses recovery. Agent processes routinely Detects exception requiring judgment Clear match Continues automatically Ambiguous Routes to human via preferred channel Human decides Agent resumes with decision ## Agent Capabilities 10 actions ### Documents 6 - **List Documents** Retrieves a paginated list of documents, filterable by collection, parent document, or template flag, with sort control. The inventory operation behind audits and migrations. - **Get Document** Retrieves one document's full details by ID: title, Markdown body, emoji, collection, author, and timestamps. The read that precedes any gated write. - **Search Documents** Full-text search across accessible documents, returning ranked matches with highlighted context snippets, optionally scoped to a collection or user. The operation that makes the wiki a mid-workflow source instead of a destination of last resort. - **Create Document** Creates a Markdown document in a collection, optionally nested under a parent, marked as a template, or published immediately. Unpublished documents remain drafts, which is how agent-authored content waits for human review. - **Update Document** Updates a document's title and body. The provided text replaces the full body by default; enabling Append adds to the end instead. Can also publish a draft. The default-replace behavior is why this page's gate exists. - **Delete Document** Deletes a document. The default moves it to trash, restorable; enabling Permanent deletes irreversibly and skips the trash. The permanent path requires a human confirmation. ### Collections 2 - **List Collections** Retrieves the workspace's collections, the top-level containers grouping related documents, paginated. The mapping step for migrations and filing logic. - **Get Collection** Retrieves one collection's details: name, description, color, icon, and permission level. ### Workspace 2 - **List Users** Retrieves workspace members, filterable by text query or membership state, paginated. Used to resolve authors and route review requests. - **Get Auth Info** Retrieves the account and workspace behind the API token, including the current user's profile and team settings. The verification step that confirms which workspace an automation is about to write into. ## Frequently Asked Questions ### What can FlowRunner do with Outline? FlowRunner agents can run List Documents, Get Document, and Search Documents in Outline, plus 7 more actions. ### Does connecting Outline to FlowRunner require OAuth? No. Outline connects to FlowRunner with an API key, no OAuth flow required. ### Can Outline trigger a FlowRunner workflow automatically? Outline doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger. **Work at Outline?** This integration exposes Outline to AI agents on every FlowRunner plan, including through MCP, at no cost to you. [See what FlowRunner offers integration partners](https://flowrunner.ai/integrations/partners), including how to keep this page current. --- Markdown version of https://flowrunner.ai/integrations/outline. Site index: https://flowrunner.ai/llms.txt