how can i add my external knowledge database link to rovo agent

[Xinhua Liao [X]]

i have a knowledge graph database using neo4j, it is hosted in external cloud and i want to add its link to Rovo, so everytime Rovo can use its knowledge as reference, how can i do that? 

thanks.

[Christos Markoulatos -Relational- · Accepted solution]

Hi @Xinhua Liao [X] ,

Currently, Rovo does not support direct database connections (like Neo4j) as a knowledge source. Rovo agents can only use:

• Atlassian products (Jira, Confluence, etc.)
• Connected SaaS apps via Rovo Connectors (Google Drive, SharePoint, Notion, Zendesk, etc.)
• Custom websites through the Custom Website Connector (Rovo crawls and indexes the site content)

This means you cannot just paste a Neo4j link and have Rovo query it natively. However, you have two practical options:

Option 1: Expose your Neo4j data as a web-accessible knowledge source

• Build a read-only web front-end or API that surfaces your graph content as HTML pages or structured docs.
• Use the Custom Website Connector in Rovo to index that site.
• Rovo will then include that content in search, chat, and agent responses.

Option 2: Build a custom integration using Forge

• Create a Forge-based Rovo Agent that calls your Neo4j API.
• This approach uses external API integration and custom actions to fetch graph data dynamically.
• You’ll need to:
  • Implement API endpoints for Neo4j queries.
  • Add a Forge function to call those endpoints.
  • Pair it with a Rovo Agent definition for natural language interaction.

Community example:
External APIs + Rovo Agents: What Actually Works

 Why no direct DB link?

Rovo is designed to respect permissions and indexing policies. Direct DB connections would bypass those controls, so Atlassian only supports connectors and APIs for now.

Helpful links

• Knowledge sources for Rovo agents
• Manage Rovo connectors
• Available Rovo connectors
• Connect custom website to Rovo
• Build a Q&A Rovo Agent with Forge

Hope the above help 😊

[Birendra Pratap Singh]

On the Atlassian Community thread you’re viewing, the guidance is clear: Rovo Agents cannot directly consume external databases or links today — including a Neo4j knowledge graph hosted in the cloud. Here’s the breakdown:

🚫 Current State
- Rovo does not support attaching external database links (like Neo4j endpoints) as a “knowledge source” it can query natively.
- There’s no configuration option to simply paste in a database URL and have Rovo use it as reference.

⚙️ Supported Approaches
1. MCP (Model Context Protocol) Tools
- Atlassian has introduced MCP servers, which allow external systems to be exposed as tools that Rovo can call.
- If you wrap your Neo4j database with an MCP-compliant server (essentially exposing queries as MCP tools), Rovo can then use those tools inside its workflows.
- This is currently the recommended path for external integrations.

2. Indirect Integration via Jira/Confluence APIs
- You can build a bridge where your Neo4j agent pushes relevant knowledge into Jira issues, Confluence pages, or comments.
- Rovo can then consume that content as part of its context when operating inside Atlassian.

3. Automation/Webhooks
- Use Jira Automation or Confluence Automation to trigger Rovo actions when your external agent posts updates.
- This way, Neo4j knowledge is surfaced indirectly through Atlassian artifacts.

📚 References
- Atlassian’s Rovo + MCP documentation is the place to watch for updates.
- The community thread you’re on confirms that direct linking of external databases is not supported yet — only via MCP or indirect automation routes .

---

👉 In short: You can’t just “add a Neo4j link” to Rovo today. The supported way is to wrap your Neo4j queries in an MCP server so Rovo can call them as tools, or push the knowledge into Jira/Confluence where Rovo already operates.

Would you like me to sketch out a step-by-step integration pattern showing how to expose Neo4j as an MCP tool for Rovo? That would give you a concrete technical roadmap.