AI Agent Tool: Confluence Knowledge Retriever

Unlock Rich Knowledge Retrieval for Your AI Agents

This n8n workflow acts as a dedicated Confluence search tool, designed to be called by your primary AI Agent workflows. When your AI Agent needs to consult your Confluence knowledge base to answer a user's question, it can delegate the search task to this connector. It takes a search query (ideally refined by an LLM in the parent workflow), queries your Confluence instance, and returns structured information including article title, link, and a content excerpt. This allows your AI Agents to provide more accurate, context-aware responses backed by your own documentation.

This tool is a key component in building AI-driven automation for Q&A, support, and internal helpdesk functions.

Key Features & Benefits

• Focused Ability: Specialized in 'knowledge retrieval' from Confluence, enhancing AI Agent capabilities.
• Seamless Integration: Designed to be triggered by parent AI Agent workflows via n8n's 'Execute Workflow' node.
• Targeted Confluence Search: Directly queries Confluence using its REST API and CQL for precise results based on the input query.
• Structured Output: Returns key information (title, link, excerpt) from the top search result, formatted for easy use by an LLM in the parent workflow.
• Knowledge Empowerment: Enables your AI Agents to tap into your existing Confluence knowledge base, significantly improving their response quality and relevance.
• Customizable Formatting: The output string can be adjusted. Note that the default output includes a sample instruction appended to the content, intended for an LLM in the parent workflow (e.g., guidance on handling password reset requests). You can modify this in the 'Return Tool Response' node.
• Modular Design: Handles the Confluence interaction, keeping your main AI Agent logic cleaner and more focused.

• An n8n instance (Cloud or self-hosted).
• Confluence instance (Cloud or Server) accessible via API.
• Confluence API credentials (e.g., API token and username/email for Basic Auth).
• A parent n8n workflow (likely an AI Agent) designed to call this tool and process its output. This parent workflow would typically handle user interaction and any LLM processing.

1. Download the n8n workflow JSON file.
2. Import the workflow into your n8n instance.
3. Configure the 'Query Confluence' (HTTP Request) node: a. Update the URL to your Confluence instance's API search endpoint (e.g., https://YOUR_DOMAIN.atlassian.net/wiki/rest/api/search). b. Set up 'Authentication' to 'Generic Credential Type', then 'Generic Auth Type' to 'HTTP Basic Auth'. c. Create or select existing n8n credentials for Confluence API access (usually your email/username and an API token).
4. Review the 'Return Tool Response' (Set) node. The default output provides title, link, and excerpt from the first search result, along with a sample instructional text for an LLM. Customize this output format and the instructional text as needed for your parent AI Agent.
5. This workflow is triggered by an 'Execute Workflow Trigger' node. It expects a query field in the incoming JSON data from the calling workflow.
6. In your main AI Agent (parent) workflow, use an 'Execute Workflow' node to call this 'Confluence Knowledge Retriever' workflow, passing the search query to it.
7. Activate this workflow so it can be called by other workflows.