Setup and configuration
Connect Your AI to Plixer in Five Minutes
Give an MCP-compatible AI client authenticated access to the network flow data Scrutinizer already collects and analyzes.
Create a Scrutinizer API token, add the Scrutinizer MCP endpoint to a supported AI client, test the connection, and ask a question about your network. Exact screens and commands vary by client.
Connect Your Client
01 Confirm the starting point
Use Scrutinizer 19.8 or later with an edition that includes MCP. Choose an MCP-compatible AI client and confirm the client can reach the Scrutinizer deployment.
02 Create the API token
A Scrutinizer administrator creates the API token for the remote AI client. The token carries permissions and roles like a user account.
03 Add the Scrutinizer connection
Enter the Scrutinizer MCP endpoint and bearer token in the AI client. Use the client’s current instructions for a remote MCP server.
04 Verify the available tools
Confirm that the client can discover the Scrutinizer tools available to the token. Start read-only, which is the default.
05 Ask a real network question
Try a bounded question with a clear time range, such as which applications used the most bandwidth on a selected link yesterday.
06 Check the result against Scrutinizer
Confirm the time range, hosts, interfaces, report output, or alarm context in Scrutinizer before expanding the workflow.
What You Need Before You Start
- Scrutinizer 19.8 or later and an edition that includes remote access to the MCP server.
- An MCP-compatible AI client.
- Network connectivity from the client to the Scrutinizer host.
- A Scrutinizer API token created by an administrator.
- The permissions required for the reports, alarms, hosts, and interfaces the workflow needs.
Frequently Asked Questions
Does Plixer AI export my network data?
No. Your network data stays in Plixer. AI accesses information through the same interface and reporting tools you use, without exporting raw flow data. Information shared with the AI model depends on your deployment configuration.
Does every user need a separate token?
Yes. Each user should have their own API token to ensure activity is logged and audited individually. Sharing tokens across teams or groups removes user-level accountability.
How do I connect my AI client to the Scrutinizer MCP server?
Setup varies by AI client. Follow our client-specific documentation to configure the MCP server URL and API token in the authentication headers.