# Confluence MCP MCP > Confluence MCP connects your AI agent directly to your organization’s knowledge base. Query technical documentation, find HR policies, and manage team resources by searching through wiki spaces using natural language prompts. It lets you pull structured data from existing pages or draft new content—like product requirement documents—and publish them without ever leaving your chat window. ## Overview - **Category:** industry-titans - **Price:** Free - **Tags:** wiki, documentation, content-creation, search-indexing, team-collaboration ## Description Think of this MCP as a direct line into your company’s entire internal knowledge base. You stop clicking tabs and start asking questions. Instead of manually navigating through different wiki spaces to find an HR policy or an old architecture decision, you just ask your AI client. It searches every relevant space, giving you the exact text snippet you need right away. Need to update documentation? Draft a new product requirement document (PRD) in chat, and then publish it directly into the correct Confluence space. You can also retrieve full pages or check who last edited something with granular detail. The real value shows up when your agent pulls knowledge from this MCP and chains that data with an external billing system MCP to automatically generate a project status report. This whole process happens securely, because Vinkius manages all credential flow through a zero-trust proxy, meaning your keys never sit on disk. It's about getting instant answers and making sure the documentation stays current—all without ever needing to log into Confluence itself. ## Tools ### add_label Adds an organizational label to any existing wiki page for easier filtering later on. ### add_page_comment Allows you to post a new comment body, formatted in HTML, onto a specific Confluence page. ### create_page Generates and publishes an entirely new wiki page into a specified space, requiring a title and content body. ### delete_page Permanently removes a Confluence page. Be aware this action cannot be undone. ### get_page Fetches the full text, metadata, and version history for one specific wiki page. ### get_page_history Retrieves a timeline of every change made to a Confluence page, showing who did it and when. ### get_space_details Pulls metadata on an entire wiki space, including its description, homepage link, and permission structure. ### list_labels Lists all labels currently applied to a specific Confluence page. ### list_page_comments Gathers every inline and footer comment from a page, including the author's name and the content of the message. ### list_pages Retrieves a paginated list of all pages within a space, optionally filtering by key. ### list_spaces Outputs a full inventory of every available wiki space in the organization's Confluence instance. ### search_confluence Searches for content across the entire wiki using complex queries targeting pages, blog posts, and comments. ## Prompt Examples **Prompt:** ``` Search Confluence for documentation regarding our 'Q3 Migration Plan'. ``` **Response:** ``` I've searched your Confluence spaces. I found two documents related to the 'Q3 Migration Plan'. The most recent one was updated two days ago in the Engineering space. Would you like a summary of its contents? ``` **Prompt:** ``` Create a new page in the 'Product' space summarizing our meeting notes from today. ``` **Response:** ``` The meeting summary has been formatted and published successfully to the 'Product' space under the title 'Meeting Notes - Today'. ``` **Prompt:** ``` List all wiki pages currently under the space key 'HR'. ``` **Response:** ``` I found 14 active pages in the 'HR' space, including 'Employee Onboarding', 'Leave Policies', and 'Corporate Benefits'. Which one would you like to review? ``` ## Capabilities ### undefined undefined ### undefined undefined ### undefined undefined ### undefined undefined ### undefined undefined ## Use Cases ### Finding an old design spec A developer needs to know the original API contract for a service. Instead of asking three people which wiki space it's in, they run `search_confluence` and immediately pull up the specific technical guideline document. ### Onboarding new hires The HR manager wants to give a new employee a full overview of company policies. They use `list_pages` on the 'HR' space key, getting a list of all relevant documents like 'Leave Policies' and 'Benefits', which they can then share. ### Project Retrospective The team needs to write up lessons learned from a past project. They use `get_page` on the final post-mortem document, gathering not just content but also metadata and version history for accuracy. ### Cleaning up old data A Product Manager notices several outdated drafts in a specific space. They can list all pages using `list_pages` and then use `delete_page` on the confirmed obsolete content, keeping the knowledge base clean. ## Benefits - Need to find a policy? Use `search_confluence` to query every piece of documentation, finding answers across spaces instead of relying on keyword matching in one area. - Drafting new content is faster than ever. After writing meeting minutes or PRDs, use `create_page` and publish them directly into the right space from your chat. - Never lose context again. If you need to know who changed a document and when, run `get_page_history` for a full audit trail of changes. - Understand your entire wiki structure with `list_spaces`. This helps map out which teams own which knowledge areas before starting a project. - The agent can pull detailed page content using `get_page`, giving you not just the text, but also space metadata and version information all at once. ## How It Works The bottom line is you use natural conversation to perform complex, multi-step administrative tasks on your corporate wiki. 1. First, authorize this MCP using your Atlassian API token. This grants your AI agent permission to interact with the wiki spaces. 2. Next, tell your agent exactly what you're looking for—for example, 'What was the decision on Project X?' The MCP executes a structured search against all connected spaces. 3. Finally, your agent returns the matching page content or metadata. You get the information presented back in plain text within your chat interface. ## Frequently Asked Questions **How do I search for documentation using `search_confluence`?** You use natural language to prompt your agent. For instance, 'Search Confluence for all documents about Q3 budget cuts.' The tool handles translating that into a complex query across the entire wiki. **Can I find out who wrote which page using `get_page`?** Yes. When you run `get_page`, it returns detailed metadata, including version history and authorship information, so you know exactly where the content came from. **What is the difference between `list_pages` and `search_confluence`?** `list_pages` gives you a list of contents within one space. `search_confluence` actually looks *inside* the content, finding matches across many spaces. **Do I need to use `add_label` every time I create a page?** No. You can call `create_page` first and then run `add_label` later if you decide the content needs further categorization or filtering. **How do I find out what spaces exist before I can use `list_pages`?** You first use `list_spaces` to get a directory of all available Confluence areas. This tool returns a list of space keys and metadata, letting you know exactly which spaces your agent can interact with next. **Should I use `get_page_history` before using `delete_page`?** Yes, always check the history first. The `get_page_history` tool retrieves every version change and timestamp for a page, giving you an irreversible audit trail of who changed what and when. **What is the best way to review comments on a specific page using `list_page_comments`?** You run `list_page_comments` and provide the target page key. It returns all inline and footer comments, including the author's name and the full content of their input. **Does `create_page` require me to know the exact format for the body content?** Yes, when running `create_page`, you must provide the title, space key, and the body content formatted in Confluence storage (HTML) format. The tool cannot guess formatting.